Main

The infant's first flavor experience may occur before birth. Recent research has revealed that the flavors of foods eaten by the pregnant woman are transmitted to amniotic fluid(1, 2). Because amniotic fluid is swallowed by the fetus, this secretion may represent the first exposure to flavors that will subsequently be provided by mother's milk(36) and then the foods of the table. In other words, breast milk may bridge the experiences of flavors in utero to those in solid foods.

For centuries, the custom of many cultures has been to introduce grain products, such as precooked cereal, as the infant's first solid food(7, 8). The cereal was sometimes prepared with mother's milk because many believed that it was essential that the food be simple, smooth in texture, and consistent in flavor(9, 10). In present times, baby food manufacturers and child care manuals advise parents to prepare the cereal with water or either mother's milk or formula, depending on the feeding regimen of the infant. Little is known about the infant's acceptance of differently flavored cereals, however. The present study investigated whether the breast-fed infants' acceptance of their first solid food, cereal, is enhanced when it is prepared with their mothers' milk and whether the infants' willingness to accept such flavored cereal is correlated with their mothers' reported willingness to try novel foods and flavors.

METHODS

Subjects. Sixteen mothers (5 primiparous, 11 multiparous) who were currently breast-feeding their infants (6 girls, 11 boys; two of the boys were fraternal twins who were tested at separate times) and had recently begun feeding them cereal were recruited from advertisements in local newspapers and from the Women, Infant, and Children Program in Philadelphia. The mothers' ages ranged from 20 to 38 y (mean = 29.4 ± 1.2 y), and the infants' ages ranged from 126 to 216 d (mean = 175.5 ± 6.8 d). During the telephone interview, mothers were queried about how long they had been feeding their infants cereal and how they prepared the cereal. Only those who had prepared the cereal with water and whose infants had been fed cereal for less than 1 mo were asked to participate in this study; the infants in the present study had been fed cereal for, on average, 12.9 ± 1.5 d. Four additional mother-infant dyads began testing but were disqualified because they did not comply with the testing procedures. All procedures were approved by the Committee on Studies Involving Human Beings at the University of Pennsylvania.

Procedures. The mothers were sent a bib, spoon, and bowl to use during the 3 d that preceded the first testing day to familiarize the infants with some aspects of the testing procedure, and they were asked to maintain feeding logs by recording the time of day they fed their infants and how much cereal their infants ate during each feed. At the same time, mothers were given a mask which covers the nose and mouth area (3M Comfort Mask; St. Paul, MN) and asked to wear this while feeding their infants. To ensure that the flavor of the milk was similar on both testing days, the mothers were instructed to eat a “bland” diet low in sulfur-containing foods and devoid of alcohol beginning 3 d before the first test session and continuing until the last test session. The mothers were given a list of foods(e.g. garlic, onion, asparagus) and spices (e.g. garlic powder, onion powder) to avoid. To encourage compliance, they were asked to record (in terms of household measures) all foods and beverages consumed during this period [see Mennella and Beauchamp(4) for a description of the methods). Mothers were also asked not to wear perfumes or use scented deodorants on each testing day.

Each infant was brought by their mother to the Monell Center on two alternate days and was last breast-fed at approximately the same time on both testing days [number of minutes since last fed: 137.7 ± 13.6versus 148.2 ± 13.7 min; t (16 df) =-1.73; p not significant]. After acclimatization to the room and personnel, the mother expressed approximately 90 mL of milk, usually from both breasts, by using an electronic breast pump (Medela, Crystal Lake, IL); the amount of milk expressed by the mothers did not differ on the two testing days[91.6 ± 10.2 versus 94.1 ± 6.1 mL; paired t(16 df) = 0.36; p not significant]. The testing session began at approximately the same time of day the infant was usually fed cereal and the type of cereal offered was the brand the infant was currently being fed (4 infants were fed Gerber Rice, 2 Gerber Oats, 4 Beechnut Rice, 3 Beechnut Oats, and 4 Earth's Best Rice cereals).

On both testing days, the mothers fed their infants using the same utensils and bib that had been used at home. Fourteen of the infants sat in a chair that had a tray attached to its arms, whereas three of the youngest infants sat in a car seat. On one testing day, the mother fed her infant the cereal prepared with mother's milk (1 part dry cereal:1 part human milk), whereas on the other testing day she fed her infant the cereal prepared with water (1 part cereal:1 part water). Nine of the infants were fed the cereal prepared with water during the first feed on the first testing day and the cereal prepared with mother's milk during the first feed on the second testing day; the order was reversed for the remaining eight infants.

The mothers were asked to wear the mask and to refrain from talking during the feeding sessions to eliminate any potential influence of the mother's facial or verbal responses on her infant's behaviors(11). Each mother was instructed to feed her infant at his or her customary pace until the infant refused the cereal on three consecutive occasions, using the criterion that the infant exhibited such behaviors as turning his or her head away, pushing the spoon away, crying, or becoming playful. The infant's behaviors were monitored by videotape, and the amount of cereal consumed by the infant was assessed by weighing the infants in their diapers only immediately before and after each feed on a Mettler PM 15 top-loading balance (Greifensee, Switzerland).

Approximately 0.5 h after the end of the first feed, the infant was again offered cereal but this time it was the other type of cereal mixture. For example, on the testing day in which the infant was first tested with the cereal prepared with water, he or she was fed the cereal prepared with mother's milk during the second feed.

Videotape analyses of infant's behaviors. Trained raters, who were unaware of the experimental conditions, scored the videotaped records of the first feeds on the two testing days by means of an IBM-based event recorder program called the Observer (Noldus, Inc., Wagenigen, The Netherlands). The behaviors were scored in real time, and the frequency of each behavioral response that occurred within a 0.1-s block, and when appropriate, the mean duration of each response and proportion of total time that the infant was engaged in the behavior was determined. Arc sine transformations of the proportional data were conducted to stabilize the variance before analysis. A separate paired t test was conducted for each of the parameters. The behaviors were defined as follows:

Infant's response to each spoonful of cereal. The rater scored whether the infant opened his or her mouth when the spoon was at a distance, when it was near the infant's mouth (within 4 cm), when it made contact with the infant's mouth, or when the infant did not open his or her mouth for each spoonful of cereal proffered.

Facial expressions. The proportion of time the infant display a neutral, positive, or negative facial expression. During scoring, the sound was turned off so that the raters would not be influenced by the infants' vocalizations.

Vocalizations. The proportion of time the infants emitted a vocalization. During scoring, the video screen was covered so that the raters would not be influenced by the infants' facial responses. Because pilot work revealed that it is difficult to distinguish between positive and neutral vocalizations, they were combined into one category.

Body movements. The proportion of time the baby was moving his or her arms, legs, or both was determined.

Questionnaires on mother's eating habits. To evaluate whether the mothers' attitudes about food and eating habits were related to their infants' willingness to accept the flavored cereal, mothers completed an 8-item scale that measured their variety-seeking tendency with respect to foods(12), a 10-item test that measures the trait of food neophobia, the reluctance to eat and avoid novel foods, and an 8-item scale that measures general neophobia, a preference for familiar situations and people versus a willingness to experience new situations and people(13). All summary statistics are expressed as the mean ± SEM.

RESULTS

Infants' feeding responses. The infants consumed significantly more cereal when it was prepared with mothers' milk compared with cereal prepared with water (p = 0.002; see Fig. 1). Moreover, although every infant exhibited signs of satiation when they were fed the cereal/water mixture, only 59% of the infants did so when fed the cereal prepared with mother's milk. Because there was no significant difference in the length of the feedings, the infants ingested the cereal/mother's milk mixture at a faster rate than the cereal/water mixture(p = 0.05; see (Table 1). This preference for the flavor of cereal made with mother's milk was also observed when cereal was again offered to the infants 0.5 h later [mother's milk versus water: 41.9 ± 11.2 versus 19.5 ± 5.59 g; pairedt (16 df) = 2.75; p = 0.014].

Figure 1
figure 1

The difference in each infant's intake of cereal flavored with mother's milk when compared with intake of cereal prepared with water. The 17 infants are rank-ordered by the amplitude of their responses. A positive score indicates the infant consumed more of the cereal prepared with mother's milk than cereal prepared with water. Therefore, 15 of the 17 infants ate more of the cereal prepared with mother's milk.

Table 1 Effects of preparing cereal with mother's milk or water on the feeding behaviors of infants

In addition to the enhanced intake, the infants displayed their preference for the cereal/mother's milk mixture (see Table 1). That is, they opened their mouths more when the spoonful of cereal was at a distance (p = 0.05) and spent less time displaying negative facial expressions (p = 0.03) and more time displaying neutral facial expressions (p = 0.03) when they were feeding the cereal prepared with mother's milk.

Questionnaires on mothers' eating habits.Figure 2 shows that there was a significant negative correlation between the mothers' food neophobia score, as determined by the questionnaire, and their infants' intake of the cereal flavored with mothers milk relative to that prepared with water [r (16 df) =-0.53; p = 0.029]. Similarly, there was a tendency for a positive correlation between the mother's variety-seeking behavior with regard to food and her infant's relative response [r (16 df) = 0.44;p = 0.080]. In other words, the mothers who exhibited a greater willingness to consume novel foods or less food neophobia had infants who consumed relatively more of the cereal prepared with mother's milk. No such correlation was observed with the mothers' general neophobia score[r (16 df) = -0.32; p not significant].

Figure 2
figure 2

Scatterplot of the infant's proportional intake of cereal prepared with mother's milk relative to total consumption (transformed by parametric analyses on left vertical axis; raw proportional scores on right vertical axis) and the mothers' food neophobia, as determined by a questionnaire. The food neophobia scores could range from 10 to 70; a lower number represents less food neophobia, a fear of novel foods.

DISCUSSION

Breast-fed infants, who had been fed cereal for approximately 2 wk but had experienced only cereal prepared with water, readily accepted the cereal when it was prepared with mother's milk. They not only consumed more of the cereal but they displayed a series of behaviors signaling their preferences for it(14). Several hypotheses, not necessarily mutually exclusive, can be suggested to explain this preference for the cereal-human milk preparation. First, the infants may be responding to differences in taste. Because human milk is sweet (approximately 7.3 g lactose/100 mL), the preference for the cereal-human milk mixture may be due to the infants' inherent preference for sweet tastes(15). Moreover, substances present in human milk, such as lactose or sodium, could mask or suppress off flavors in the cereal thereby enhancing its palatability(16, 17).

Second, the infants may have a mechanism for estimating the caloric value of a food. For example, if there is a sensory mechanism by which the fat content can be determined, the infants' preference for the cereal prepared with mothers' milk could reflect this ability to identify the higher caloric substance.

Third, the volatile flavor components of human milk may heighten the acceptability of foods which contain it. We have shown that rather than being a bland flavorless substance, human milk is rich in flavors, which directly reflect the dietary choices of the mother, and that both breast- and bottle-fed infants alter their pattern of suckling when feeding flavored milk or formula, respectively(36). Thus cereal with breast milk may be more palatable because it contained flavors that infants find attractive either because of their inherent pleasantness or because of previous experiences with these flavors in amniotic fluid(1) or human milk(36).

Fourth, the addition of breast milk to cereal may have increased its familiarity or reduced its neophobia which, in turn, heightened acceptance(18). Mother's milk is a highly preferred food that is very reinforcing for young mammals(19). Therefore, although these infants had never experienced cereal prepared with breast milk, breast milk was an extremely familiar and highly attractive food, whereas cereal was still relatively unfamiliar to the infants. This contention is supported by the finding that the infants consumed more of the cereal prepared with mother's milk on the testing day in which they had first satiated on the cereal-water mixture.

Finally, and in direct contrast to the fourth hypothesis, the infants may have been attracted by the novelty of the cerealbreast milk flavor relative to the cereal-water preparation to which they were familiar. Novel stimuli can be arousing, thereby stimulating activities including eating(20). That the mothers who exhibited a greater willingness to consume novel foods or less food neophobia, as determined from questionnaires, had infants who consumed relatively more of the cereal prepared with mother's milk might suggest that there could be a genetic basis underlying willingness to sample novel flavors and foods(21). A similar findings has been obtained in the formula-fed infants' acceptance of a novel, protein hydrolysate formula(22). This relationship between mothers' reported eating habits and their infants' response to flavors could be based on anything from genetic effects on flavor sensitivity(21, 23, 24) to more general effects on the response to novelty. Because we found no correlation in either study between the infants' relative intake and the mothers' general neophobia score, the relationship was apparently related to the mothers' dietary habits and not to a more generalized phenomenon such as novelty-seeking behavior.

This relationship between the mothers' variety-seeking behaviors and their infants' relative acceptance of the cereal flavored with mother's milk may also be a consequence of preand postnatal experiences. It is possible that mothers who were more willing to consume variety and novelty in foods during pregnancy and lactation exposed their infants to a more varied sensory experience, because a variety of flavors consumed by the lactating woman, and at least one flavor consumed by pregnant women, garlic, readily passes into milk(36) and amniotic fluid(1), respectively. There is evidence in other animals that prenatal exposure to flavors can affect later preferences for these flavors(25) and that postnatal exposure to a variety of flavors in mother's milk renders the young mammal more likely to accept novel flavors during weaning(26).

The meaning of the word weaning is rather ambiguous. As Dana Raphael(8) has noted, cultures around the world define the term a variety of ways, definitions that include the time in an infant's life when complementary foods or fluids, even water, are introduced in addition to breast milk or formula or the time when the method of delivering the food changes (e.g. breast to bottle, bottle to cup, cup to spoon). Non-food-related definitions are also seen in cultures that define weaning in terms of certain rites of events in the infant's life, such as when the infant teethes, reaches a certain age, or when the haircutting ceremony is performed. In contrast, others have described weaning as a major transformation spanning an extensive period of ontogeny during which mothers gradually introduce their infants to the foods indigenous to their group or culture(8, 27).

The root origin of the word “weaning” also varies from culture to culture. For example, in German the verb is entwöhnen and in Danish, vanne fra. Both are related to the Old Englishweanian, “to accustom,” which is a contraction ofwee ane (wee one, meaning “small one”). Both the Danishvanne and the German wohnen describe the breaking of a habit, but the same root is also used to describe the building of a habit(vanne til in Danish, gewohnen in German). From the perspective of flavor, the root definitions of “weaning” that mean“to accustom” or “to build a habit” may actually describe what parents around the world do when they gradually introduce the flavors which have come to identify their culture. Moreover, if weaning is considered that process by which an infant is accustomed to the foods and flavors of his or her culture, then we suggest that this process extends even longer, beginning in utero and continuing throughout lactation, when the infant becomes accustomed to the mother's diet via the transmission of food volatiles to amniotic fluid and human milk, and extending throughout childhood.

Because human milk is rich in flavors that directly reflect the dietary choices of the mother(36), the transition from a diet exclusively of human milk to a mixed diet may be facilitated by providing the infant with bridges of familiarity such that the infant experiences a commonality of flavors in the two feeding situations. In this case, human milk would be familiar and would facilitate the introduction of a novel food, cereal. Nonhuman animal studies have documented this phenomenon(28, 29), and in fact, it has been exploited in the development of weaning foods for pigs(30). Whether such a transition would be different for the infant who experiences a monotony of flavor in standard formulas warrants further study(4, 6).