The Pennsylvania State University. The Graduate School. Department of Human Development and Family Studies AN EXAMINATION OF TEMPERAMENTAL

Transcription

1 The Pennsylvania State University The Graduate School Department of Human Development and Family Studies AN EXAMINATION OF TEMPERAMENTAL APPROACH/WITHDRAWAL AND RESPONSES TO NOVEL FOODS ACROSS INFANCY AND EARLY CHILDHOOD A Dissertation in Human Development and Family Studies by Kameron J. Moding 2016 Kameron J. Moding Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy May 2016

2 The dissertation of Kameron J. Moding was reviewed and approved* by the following: Cynthia A. Stifter Professor of Human Development and Psychology Dissertation Adviser Chair of Committee Leann L. Birch Distinguished Professor Emeritus of Human Development and Family Studies Kathleen L. Keller Assistant Professor of Nutritional Sciences and Food Science Douglas M. Teti Professor of Human Development and Family Studies, Psychology, and Pediatrics Department Head, Human Development and Family Studies Eva S. Lefkowitz Professor of Human Development and Family Studies Professor-in-Charge, Graduate Program *Signatures are on file in the Graduate School

3 iii ABSTRACT Food neophobia emerges during the transition to solid foods during infancy and refers to the tendency to reject novel or unknown foods. This rejection of novel foods has been shown to have adverse effects on the diets of infants and young children. For example, infants who reject new foods may not be able to consume the variety of foods necessary for adequate nutrition. Similarly, food neophobia during early childhood has been linked to poor dietary outcomes, such as less dietary variety and lower consumption and liking of vegetables. For these reasons, it is important to identify factors that may contribute to food neophobia during the first years of life. The two studies of this dissertation project investigated temperamental differences in the tendency to approach or withdraw from novel stimuli as one such factor. Temperamental approach/withdrawal processes have been extensively examined in response to new toys, new people, and new situations, but they have rarely been examined in response to novel foods. Thus, the overarching goal of this dissertation was to examine whether temperamental approach/withdrawal tendencies drive infants and young children s reactions to novel foods. Data for both studies were drawn from a longitudinal study following 115 infants and their families across the first five years of life. Latent profile analysis was used to identify subgroups of children who exhibited similar patterns of approach/withdrawal and affective responses to novel stimuli during toddlerhood. Infants responses to novel toys and novel foods were examined as predictors of the resulting temperament groups (inhibited, exuberant, and balanced) in Study 1. In Study 2, the toddler temperament groups were used as predictors of food neophobia at 4.5 years of age.

4 iv The goal of Study 1 was to investigate whether infants approach/withdrawal tendencies underlie their responses to novel foods. This goal was addressed by examining whether infants acceptance and rejection of novel foods were concurrently and longitudinally related to well-established measures of temperamental approach/withdrawal, such as responses to novel toys and objects. As predicted, infants who exhibited high levels of approach in response to the novel toys at 12 months also tended to show high levels of acceptance and low levels of rejection in response to the novel foods concurrently. Conversely, infants responses to the novel toys at 6 months did not consistently predict their responses to the novel foods at the same age. Together, these findings suggest that approach/withdrawal behaviors are only consistent across novel contexts after infants develop the ability to inhibit their approach tendencies and once individual differences in wariness come online in the second half of the first year of life. Furthermore, infants responses to the novel foods at 12 months, but not 6 months of age, predicted their temperamental style during toddlerhood. Infants who showed higher levels of rejection in response to the novel food at 12 months exhibited fewer approach behaviors in an unfamiliar environment with an experimenter and were more likely to be classified as inhibited compared to exuberant at 18 months of age. In sum, the results of Study 1 were largely consistent with our proposal that infants reactions to novel foods are likely driven by underlying temperamental approach/withdrawal processes. Study 2 examined whether temperamental approach/withdrawal processes concurrently and longitudinally predicted parent ratings and behavioral observations of food neophobia at 4.5 years of age. Additionally, this study explored whether parent characteristics and feeding styles moderated the relationship between toddler

5 v approach/withdrawal and food neophobia during preschool. As expected, parent ratings and behavioral observations of temperamental approach/withdrawal were associated with concurrent measures of food neophobia at 4.5 years of age. Similarly, temperamental withdrawal during toddlerhood, specifically high levels of negative affect, low levels of positive affect, and the inhibited temperament type, predicted food neophobia over three years later. Collectively, the results of this study emphasize that children who are temperamentally inhibited are at a heightened risk for developing high levels of food neophobia during childhood. Furthermore, evidence emerged to show that these neophobic tendencies in inhibited children were exacerbated by a maternal pressuring feeding style. In conclusion, this dissertation adds to the existing literature on food neophobia by revealing that temperamental approach/withdrawal processes were linked to food neophobia during infancy and early childhood. The results across both studies also contribute to the literature on approach/withdrawal processes by confirming that novel foods are another type of novel stimulus that elicits individual differences in approach/withdrawal behaviors. Future work should continue to examine parenting practices and environmental influences that may decrease neophobic tendencies in temperamentally inhibited children, who are most at-risk for developing high levels of food neophobia during childhood.

6 vi TABLE OF CONTENTS List of Figures... viii List of Tables... ix Acknowledgements... xi Chapter 1 Introduction... 1 Food Neophobia... 2 Repeated exposure studies... 3 Conceptualization and measurement of food neophobia... 5 Individual differences in neophobia Temperament Approach/withdrawal processes Measurement of temperament Temperament/personality and food neophobia The Present Dissertation The Back to Baby Basics (B2BB) Project References Chapter 2 Study 1: Novel Food: Another Stimulus that Elicits Infant Approach/Withdrawal Behaviors Predictors of toddler temperamental style The Present Study Methods Participants Procedures Measures Data analysis plan Results Preliminary analyses Primary analyses Discussion Conclusions References Chapter 3 Study 2: Examining Temperamental Approach/Withdrawal as a Concurrent and Longitudinal Predictor of Food Neophobia during Early Childhood The Present Study Methods Participants

7 Procedures Measures Data analysis plan Results Preliminary analyses Primary analyses Discussion Conclusions References Chapter 4 Conclusion Implications Study limitations and future directions References vii

8 viii LIST OF FIGURES Figure 2.1. Graph of means for Latent Profile Analysis (LPA) indicators by temperament group Figure 3.1. Graph of means for Latent Profile Analysis (LPA) indicators by temperament group

9 ix LIST OF TABLES Table 2.1. Definition of Codes and Reliability for Infants Responses and Maternal Responsiveness during the Novel Food Tasks at 6 and 12 Months of Age Table 2.2. Definition of Codes and Reliability for Infants Responses and Maternal Responsiveness during the Novel Toy Tasks at 6 and 12 Months of Age Table 2.3. Model Fit Indices for the 18-Month Temperament Group Latent Profile Analyses Table 2.4. Descriptive Statistics for Covariates, Infants Responses to the Novel Toys and Foods, and Toddler Temperament Variables Table 2.5. Correlations between the Infant Novel Food, Infant Novel Toy, and Toddler Risk Room Responses Table 2.6. Multiple Regression Results Predicting Infants Novel Food Responses from their Novel Toy Responses at 6 Months Table 2.7. Multiple Regression Results Predicting Infants Novel Food Responses from their Novel Toy Responses at 12 Months Table 2.8. Multiple Regressions Predicting Toddlers Responses to Novelty from their Novel Toy Responses Table 2.9. Multiple Regressions Predicting Toddlers Responses to Novelty from their Novel Food Responses Table Multinomial Logistic Regressions Predicting Temperament Groups from Infants Novel Toy and Food Responses Table 3.1. Model Fit Indices for the 18-Month Temperament Group Latent Profile Analyses Table 3.2. Descriptive Statistics for Covariates, Temperament, Parenting, and Food Neophobia Variables Table 3.3. Correlations between Covariates, Temperament, Parenting, and Food Neophobia Variables Table 3.4. Multiple Regression Results Predicting Behavioral Neophobia from Toddler Temperament Composite Variables Table 3.5. Multiple Regression Results Predicting Parent-Rated Neophobia from Toddler Temperament Composite Variables Table 3.6. Multiple Regression Results Predicting Behavioral Neophobia from Toddler Temperament Group Classifications

10 Table 3.7. Multiple Regression Results Predicting Parent-Rated Neophobia from Toddler Temperament Group Classifications x

11 xi ACKNOWLEDGEMENTS First, I would like to thank my dissertation chair and advisor, Dr. Cindy Stifter. You are an incredible mentor, and I am extremely grateful for the opportunity to work with you throughout graduate school. You have taught me the ins and outs of temperament theory and challenged me to think deeply about every aspect of my research. Thank you for constantly pushing me to become a better writer. I am thankful for your endless guidance and support. I would also like to thank Dr. Leann Birch who served on every one of my committees. I hope you know that you have played an integral role in the development of my research interests and plans for the future. Thank you for your wisdom, encouragement, and support over the last several years. You are an amazing role model. I want to thank my other committee members, Dr. Doug Teti and Dr. Kathleen Keller. Doug, thank you for always challenging me to think critically. I am grateful for the knowledge you have shared with me on parenting and infant attachment. As I move forward with my own research, I will continue to incorporate these concepts. Kathleen, thank you for your guidance on my committees and through the Childhood Obesity Prevention Training (COPT) program. You are an incredible resource, and I know that I will continue to seek your professional development advice as I navigate my own career path. I would also like to express my appreciation to the families who participated in the Back to Baby Basics (B2BB) project. Also, thank you to the many undergraduate research assistants who helped with data collection, behavioral coding, and data entry for our project. The overall study, as well as this dissertation, would not have been possible without you. I would also like to thank the COPT program for funding the majority of my time in graduate school and my travel to conferences. This work was funded by a grant from the National Institutes of Digestive Diseases and Kidney (award number DK081512).

12 xii I am incredibly grateful for my cohort, friends, and labmates. I will never be able to fully express how thankful I am for your constant support and encouragement. I am confident that I would not have made it through this journey without you. Thank you to my family for your unwavering love and support. Thank you to my brother, Everett, for being the best role model I could ever ask for. Your accomplishments are truly inspiring. Finally, I want to thank my parents and family friends for always believing in me. I appreciate your support and encouragement more than you will ever know.

13 1 Chapter 1 Introduction Humans, like other omnivores, are faced with a dilemma when trying novel foods: they must accept the new foods to ensure dietary variety, but they also risk the possibility of ingesting toxic substances. In order to protect themselves from harm, humans have the tendency to reject novel or unknown foods, which is known as food neophobia (Birch, 1999; Dovey et al., 2008; Rozin, 1976). This tendency is considered adaptive because it prevents humans from ingesting potential poisons, but it also interferes with the acceptance of novel foods which is necessary for adequate nutrition. Recent research has revealed individual differences in infants and young children s neophobic tendencies and responses to novel foods (Forestell & Mennella, 2012; Moding, Birch, & Stifter, 2014). Neophilic individuals tend show positive responses to novel foods, whereas neophobic individuals tend to reject new foods (Pliner & Hobden, 1992), but, there is surprisingly little research investigating possible explanations for these differences. The present dissertation proposes one possible factor that may account for these individual differences in food neophobia: temperament, or differences in emotional and behavioral reactivity and regulation (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981). One aspect of temperament, in particular, is the individual s behavioral tendency to approach or withdraw from novel stimuli (Kagan & Fox, 2006; Rothbart & Bates, 2006), which may account for individual differences in reactions to novel foods. The present dissertation will investigate whether temperamental approach/withdrawal underlies food

14 2 neophobia or responses to novel foods through three primary aims: 1) examining whether infants and young children s responses to novel toys or objects concurrently predict their responses to novel foods; 2) investigating whether infants responses to novel toys and foods predict their temperamental style during toddlerhood; and 3) examining whether toddler temperamental style can predict food neophobia during early childhood. This introduction provides the basis for this dissertation project and will be divided into four primary sections. The first section will introduce the construct of food neophobia and highlight issues with current conceptualizations and measurement in young children and infants. The second section will present modern temperament theories and current measurement strategies for studying temperament. The third section will discuss the limited research that examines the relationship between temperament or personality and food neophobia. The fourth and final section of this introduction will discuss the overarching aims of this dissertation and the specific aims for each of the two dissertation studies. Food Neophobia Food neophobia refers to the tendency to reject novel or unknown foods (Birch, 1999; Dovey et al., 2008). This concept originated from research with non-human animals. Omnivores, such as rats and monkeys, do not readily accept new foods. Instead, they tend to show wariness and avoidant behaviors when presented with a novel food. These types of responses serve a protective role by preventing animals from ingesting harmful toxins (Rozin, 1976). Since humans are omnivores, it is believed that they also

15 3 exhibit neophobic responses to novel foods. These responses may serve the same adaptive function in humans as they do in animals: to prevent the ingestion of toxic substances (Birch, 1999). However, these responses, while adaptive, are at odds with the acceptance of novel foods. Thus, researchers have primarily focused their efforts toward identifying strategies that reduce neophobic responses and increase acceptance of novel foods in children and infants, such as repeated exposure to the novel food over time. Repeated exposure studies Very early work on the development of food preferences in young children identified the degree of familiarity as one of the most salient dimensions of a food that predicts children s food preferences. In one study, children s liking for a variety of fruits were driven by two latent dimensions: familiarity and sweetness. Familiarity emerged as the most important dimension predicting preferences in three-year-olds; in four-year-olds, sweetness was the most important dimension, followed by familiarity. This early work on food preferences suggests that children tend to prefer and like foods that are familiar and dislike foods that are novel (Birch, 1979). Based on these initial findings, researchers began to focus on strategies that would help children overcome the obstacle of unfamiliarity in order to accept new foods. One such strategy is repeated exposure to novel foods overtime. Many research studies have confirmed that when children are given multiple opportunities to taste a novel food, their preference for that food increases (e.g. Birch & Marlin, 1982; Birch et al., 1987). Additional research has emphasized that this result is specifically due to exposure to the

16 4 taste of the food; other types of exposures, such as visual exposures, are not sufficient to increase acceptance of a novel food (Birch et al., 1987). The success of repeated exposures has been replicated in multiple samples of young children, ranging in age from 2 through 6 or even 8 years of age, (Anzman-Frasca, Savage, Marini, Fisher, & Birch, 2012; Birch & Marlin, 1982; Wardle et al., 2003) and research has suggested that as many as eight to ten exposures are required to increase acceptance of novel foods in this age group (Birch et al., 1987; Sullivan & Birch, 1990). Thus, as children become familiar with a new food, their acceptance of the food increases. After repeated exposure emerged as a successful strategy to decrease food neophobia in children (Birch et al., 1987; Birch & Marlin, 1982), researchers began to test whether repeated exposures can also increase acceptance of novel foods in infants. However, it is important to note that before infant repeated exposure studies began, infants responses to novel foods had not been well-studied; there was a lack of research investigating whether or not infants show neophobic or negative responses to novel foods. Previous research had revealed that infants exhibit differential facial expressions in response to the four basic tastes (i.e. sweet, salty, sour, and bitter) (Rosenstein & Oster, 1988), but differential facial expressions in response to novel foods had not been examined. Further, there was no existing research to compare whether infants differ in their consumption of novel versus familiar foods. Despite this lack of research, researchers began to test whether the success of repeated exposures would extend down to infancy. In the first repeated exposure study in human infants, Sullivan and Birch (1994) showed that infants increased their consumption of a novel food after exposure to the

17 5 food over a 10-day period. The increases in consumption were particularly strong for breastfed infants who showed greater intake of the novel food after repeated exposures compared to formula-fed infants. Birch, Gunder, Grimm-Thomas, and Laing (1998) furthered this work by investigating the amount of exposures necessary to increase consumption of an initially novel food. They found that only a few exposures were required: infants in their study doubled their intake on only the second feeding of a novel food. This result contrasts with the results in young children who require multiple exposures before increasing their consumption of a novel food (Birch et al., 1987; Birch & Marlin, 1982). Weaker neophobic tendencies may be adaptive in infancy when parents control their children s access to food, whereas stronger neophobic responses may be more adaptive in older children who are more mobile and independent (Birch et al., 1998). Since these original studies, many studies have replicated the result that repeated exposure increases acceptance and consumption of novel foods in infants (Maier, Chabanet, Schaal, Issanchou, & Leathwood, 2007; Remy, Issanchou, Chabanet, & Nicklaus, 2013). Conceptualization and measurement of food neophobia Although it has been shown that neophobia decreases with repeated exposures in both children and infants, there is surprisingly little research on what constitutes an initial neophobic response. In other words, there are few studies characterizing the expression of neophobia and no studies describing how this expression may be similar or different in infancy versus early childhood. Based on the results of separate research studies with

18 6 infants (Birch et al., 1998; Sullivan & Birch, 1994) and children (Addessi, Galloway, Visalberghi, & Birch, 2005; Birch et al., 1987; Wardle et al., 2003), researchers believe that neophobic reactions start at a low level during infancy, reach a peak during early childhood, and decline into later childhood and adolescence. This described trajectory suggests that the construct of food neophobia is present from infancy through childhood, but there are no existing longitudinal studies tracking its development. Without this type of research, it is difficult to tell whether reactions to novel foods during infancy and early childhood are driven by the same underlying construct. Further, there has been very little, if any, research to demonstrate whether neophobia is expressed differently in infants versus older children. Food neophobia in children Food neophobia in children is typically defined as unwillingness to eat, or avoidance of, novel foods (Birch & Fisher, 1998; Dovey et al., 2008). Young children are aware when a food is novel, so they often reject new foods based on visual appearance or smell rather than taste (Dovey et al., 2008). Since food neophobia serves to protect individuals from ingesting harmful substances, some researchers believe that food neophobia can only be present until a novel food enters an individual s mouth for the first time. After this point, the food is no longer novel and the individual has already been exposed to the substance; thus, neophobia can no longer serve its protective role. Further, some researchers believe that any responses to the food that occur after tasting it, such as

19 7 acceptance or rejection of the food s flavors, fall under the umbrella of picky eating instead of neophobia (Dovey et al., 2008). Food neophobia, in the form of reluctance to taste novel foods, has been captured in young children using both observational methods and parent reports. The most widely used observational paradigm was developed by Pliner and colleagues (Pliner, 1994; Pliner & Loewen, 1997) and involves the presentation of a variety of novel food items to children in a laboratory. The children are asked if they would be willing to taste each food. Familiar food items are also presented in order to assess the children s general willingness to taste any food in an unfamiliar environment. Children s selections for all foods are recorded before the children taste any foods in order to prevent later choices from being affected by hedonic responses to earlier foods. A behavioral food neophobia score is typically calculated as the relative ratio of willingness to taste novel versus familiar food items (Pliner & Loewen, 1997). Acceptance and liking of the novel foods after they are tasted are not included as part of the behavioral neophobia measure. Food neophobia in children has also been assessed via parent-report questionnaire. The most widely used instrument is the Food Neophobia Scale for Children (FNS-C) which was adapted from Pliner and Hobden s (1992) Food Neophobia Scale for adults. The FNS-C consists of six items assessing children s general willingness to taste novel foods (e.g. My child does not trust new foods ; My child is afraid to eat things s/he has never had before ). The final parent-reported neophobia score is created by averaging the parents reports across the six items. Higher scores indicate greater food neophobia.

20 8 Both observational and parent-report measures of food neophobia in children have strengths and weaknesses. For example, one strength of laboratory or observational measures of food neophobia is that they are based on children s actual behaviors when presented with novel foods that are not affected by reporting bias (i.e. inaccurate recall or parents own levels of neophobia). However, observational measures have weaknesses such as the limited number of foods that can be presented to the children and that children s responses to the foods may be affected by the unfamiliar environment, experimenter, or current hunger levels. These limitations do not exist with parents reports of their children s food neophobia, as parents see their children s responses to a variety of novel foods. However, parent reports are affected by reporting bias and may not be able to capture within-person variability in a child s responses to novel foods. For example, children may be very willing to taste some novel foods, but not others. With questionnaire measures of neophobia, parents must consider their children s responses to a variety of novel foods and select just one score for each item on the rating scale. Although the number of studies utilizing both observational and parent-report measures of neophobia in children is limited, there is some evidence for convergent validity between the two types of measures. In one study, children s levels of behavioral neophobia were significantly and moderately correlated with parent reports of food neophobia. Further, parents predictions of whether or not their child would taste each novel food were fairly accurate and accounted for a significant portion of the variance in children s willingness to try the new foods (Pliner, 1994). Although this preliminary evidence suggests that the observational and parent-report measures of neophobia are related, more research is needed in order to establish convergent validity.

21 9 Food neophobia in infants The conceptualization of food neophobia in children as unwillingness to taste novel foods and the belief that neophobia exists only until the food enters an individual s mouth (Dovey et al., 2008) are problematic when applying the construct of food neophobia to infants. Unlike children, infants are typically unaware that they are about to taste a new food. Most foods, whether they are familiar or novel, are presented to infants on a spoon. Thus, recognition that a food is novel is likely to occur only after the new food enters an infant s mouth for the first time. If neophobia can only exist until the point at which a novel food enters an individual s mouth, infants would not be able to experience food neophobia. Following this line of reasoning, some researchers have avoided using the term food neophobia with infants (Dovey et al., 2008). The widely used observational paradigm assessing children s willingness to taste novel foods is also problematic for use with infants. Since infants do not know when a food is novel, it does not make sense to assess their behaviors before the food enters their mouths for the first time. However, if a broader definition of food neophobia is utilized that includes infants initial reactions to a food once they recognize it as novel (i.e. after it enters their mouths for the first time), it may be possible to measure neophobia in infants. In fact, three measures have been used with infants to show that their responses to novel foods may be somewhat negative indicating the possible presence of neophobia. These measures are consumption of the food (in grams), facial expressions, and behaviors (i.e. acceptance and rejection).

22 10 First, infants show a less favorable response to novel compared to familiar foods when consumption is considered. More specifically, infants consume less of a novel food on the first exposure compared to later exposures of the same food (Birch et al., 1998; Sullivan & Birch, 1994). Second, infants facial expressions when tasting a novel food also appear to be somewhat negative. Forestell & Mennella (2007) showed that the majority of infants display facial expressions indicating distaste, such as squints, brow movements, and upper-lip raises when tasting a novel vegetable (i.e. green beans). These types of facial expressions have also been linked to lower ratings by the infants mothers on how much they thought their infant enjoyed the novel food (Forstell & Mennella, 2012). Third, infants often exhibit behaviors indicating rejection (i.e. swatting the spoon, turning away, or fussing/crying) when tasting a novel food for the first time. In one study, the mean ratio of infants rejection behaviors compared to all behaviors exhibited while tasting a novel food (i.e. rejection, positive, or neutral behaviors) was.40 indicating a high level of rejection during the first exposure to a novel food (Moding et al., 2014). Taken together, the results from these studies suggest that infants may show a neophobic or negative response to novel foods when their initial responses and behaviors following the first taste of a novel food are considered. Individual differences in neophobia Depending on how food neophobia is defined, there is evidence to suggest that children and infants show neophobic responses to novel foods. However, research has also found individual differences in food neophobia for both age groups. During infancy,

23 11 the factors affecting neophobia are often food-related, such as maternal diet during pregnancy, breastfeeding, and exposure to a variety of novel foods. In older children, individual differences in neophobia continue to be affected by environmental factors surrounding food, such as parent feeding practices (Blissett & Fogel, 2013). However, it is important to note that a few studies indicate that individual differences in child neophobia are partially due to genetic factors (Cooke, Haworth, & Wardle, 2007; Faith, Heo, Keller, & Pietrobelli, 2013). An infant s feeding history appears to play a role in his or her acceptance of novel foods at weaning. This feeding history begins prenatally, continues through lactation, and includes the transition to solid foods. For example, maternal diet during pregnancy contributes to infants later acceptance of novel foods because many flavors from a mother s diet are transferred through the amniotic fluid. The developing organism can taste and become familiar with these flavors, which may affect its later acceptance of these flavors in solid foods (Mennella, 2014; Mennella, Jagnow, & Beauchamp, 2001). In one study, mothers who drank carrot juice during pregnancy had infants who showed fewer negative facial expressions when tasting a carrot-flavored cereal for the first time compared to infants of mothers who drank water during pregnancy. Further, these same infants were rated by their mothers as enjoying the carrot-flavored cereal more than plain cereal (Mennella et al., 2001). These results suggest that early exposure to flavors in the womb may affect infants later acceptance of novel foods with the same flavors. These results are also consistent with the animal literature suggesting that animals tend to prefer flavors they experienced during prenatal development (Hepper, 1988).

24 12 Another aspect of an infant s feeding history that may contribute to individual differences in novel food acceptance is exposure to breast milk. However, there are mixed results on this topic. One study found no differences between breastfed and formula-fed infants on their initial intake of a novel food (Sullivan & Birch, 1994), whereas another study showed that breastfed infants ate more calories of a novel fruit (i.e. peaches) than did formula-fed infants (Forestell & Mennella, 2007). Differences in results between the studies may be attributed to differences in mothers consumption of the target novel food while lactating. Similar to how flavors from a mother s diet are transmitted to the amniotic fluid, some flavors are also transmitted through a mother s breast milk to the infant (Mennella, 2014; Mennella et al., 2001). Therefore, breastfeeding may only offer an advantage to infants if mothers consume the novel food while lactating (Forestell & Mennella, 2007). Additional research on the relationship between breastfeeding and infants acceptance of novel foods is needed to confirm when and how breastfeeding can offer an advantage for infants as they begin the transition to solid foods. A final aspect of an infant s feeding history that leads to individual differences in acceptance of novel foods is whether the infant has been fed a variety of solid foods. For example, Gerrish and Mennella (2001) and Lange et al. (2013) showed that previous exposure to a variety of foods led to increased acceptance of a novel food on the first exposure. Similarly, Mennella, Nicklaus, Jagolino, and Yourshaw (2008) found that repeated exposure to a variety of fruits increased acceptance of a novel fruit, but not a novel vegetable. Finally, Maier, Chabanet, Schaal, Leathwood, and Issanchou (2008) found that offering three vegetables for three consecutive days each did not increase

25 13 consumption of a new vegetable (i.e. peas), whereas alternating the three vegetables daily did increase consumption of the new vegetable. Together, these results suggest that exposure to a variety of flavors may lead to increased acceptance of new foods. However, this effect may be dependent on the similarity between the variety of foods and the target novel food (Mennella et al., 2008; See Birch et al., 1998 for a similar finding) and how often the various foods are rotated. Research with children has also highlighted a number of environmental factors that predict individual differences in neophobia. In particular, findings suggest that parenting, both parent feeding practices and behaviors, is associated with neophobia in children. For example, parental or adult modeling of novel food consumption is related to children s acceptance of novel foods. Addessi et al. (2005) found that two to five-yearold children consumed more of a novel food when an adult model was also eating the same food at the same time. Similarly, Harper and Sanders (1975) found that children were more likely to accept a new food when adults were also eating the new food compared to when adults only offered the food to the child. Further, the youngest children in the study (ranging in age from 14 to 20 months) accepted the novel foods even more frequently when the foods were offered by their mother compared to another adult. These results suggest the importance of adult models, particularly parents, in toddlers and young children s willingness to taste novel foods. Another factor that may be associated with individual differences in children s neophobia is the parents own levels of neophobia. Parents who have high levels of food neophobia tend to have children who also have high levels of neophobia (Falciglia, Pabst, Couch, & Goody, 2004; Galloway, Lee, & Birch, 2003; Pliner & Loewen, 1997).

26 14 However, it is important to note that since parents typically report on their own and their children s neophobia, the relationship between parent and child neophobia may actually be due to shared or common method variance, rather than a true relationship (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). In support of this argument, one study showed that parents ratings of their own neophobia were correlated with their ratings of their children s neophobia, but behavioral ratings of maternal and child neophobia were uncorrelated (Pliner & Loewen, 1997). The findings that parents and children s levels of neophobia are positively associated may be indicative of a true relationship that may be accounted for by genetics. In one study, genetic factors accounted for 72% of the variance in neophobia for children ages 4 through 7 (Faith et al., 2013). This estimate is comparable to estimates of heritability in adult populations (Knaapila et al., 2007). However, there is also evidence that the relationship between parent and child neophobia may be due to environmental factors. Parents who consume a varied diet tend to have children with lower levels of neophobia (Faith et al., 2013; Galloway et al. 2003). If parents have an unvaried diet due to their levels of neophobia, it is possible that they may not offer a variety of foods to their children which may subsequently affect their children s levels of neophobia. More research is needed to confirm whether the relationship between parents and children s levels of neophobia is due to genetic or environmental factors; however, regardless of the explanation, parents own levels of neophobia appear to be associated with individual differences in their children s willingness to try novel foods. Taken together, the existing research on food neophobia suggests that there are many factors, most of which are related to parenting behaviors and decisions around

27 15 food, which contribute to individual differences in food neophobia. However, there is one additional factor that may predict levels of food neophobia in infants and children that has been largely understudied in food neophobia research. This factor is the child s temperament, or individual differences in emotional and behavioral reactivity and regulation (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981). Temperament Temperament is a rubric of inherent traits that drive human behavior. Differences in temperament appear early in life and lead individuals to respond differently to the same stimuli (Goldsmith et al., 1987). Researchers who study temperament have developed theoretical models that share several points of consensus. First, researchers agree that temperament refers to individual differences in behavioral tendencies, as opposed to differences in specific behaviors. Second, temperament traits are proposed to be strongly biologically- or constitutionally-based. Finally, temperament traits are believed to be relatively stable, but the behavioral expression of these traits can be modified through interaction with the environment. While temperament theorists generally agree on these points, they disagree somewhat on the dimensions and boundaries of temperament (Goldsmith et al., 1987). Some theorists believe that temperament should only include dimensions that are personality traits (Buss & Plomin, 1975) or primary emotions (Goldsmith & Campos 1982, 1986), whereas other researchers include a much wider range of dimensions (e.g. high intensity pleasure and duration of orienting) (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981).

28 16 Researchers have developed many models that incorporate each of the basic characteristics of temperament described above. Two developmental temperament models that are particularly relevant to this dissertation project are the models proposed by Thomas and Chess (Chess & Thomas, 1973; Thomas & Chess, 1977) and Rothbart (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981). The model developed by Thomas and Chess was one of the first modern temperament theories (Chess & Thomas, 1973; Thomas & Chess, 1977). Based on extensive parent interviews, Thomas and Chess described temperament as stylistic attributes of behavior that appear during early infancy and are exhibited in response to external stimuli. They focused specifically on differences in how, rather than why, individuals expressed behaviors (Chess & Thomas, 1973). Thomas and Chess model is important for the present study because it was the first developmental model to feature the dimension of approach/withdrawal in children. This dimension characterizes an individual s response to novel stimuli making it a central temperamental characteristic of interest in the present studies. Thomas and Chess were also the first researchers to suggest that approach/withdrawal processes may be elicited in response to novel foods (Thomas & Chess, 1977). This idea was not a central aspect of their theory, but they suggested that a new food is one stimulus that should elicit approach/withdrawal behaviors. Finally, Thomas and Chess model served as the basis for many contemporary models and assessments of temperament (Rothbart & Bates, 2006). Rothbart and colleagues (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981) expanded on Thomas and Chess (1977) model to propose a second approach to temperament that guides the present studies. Rothbart defines temperament as

29 17 constitutionally-based individual differences in reactivity and self-regulation. Reactivity encompasses a broad range of behavioral and physiological responses to environmental stimuli that can be measured by parameters such as latency, duration, and intensity. Regulatory processes, on the other hand, serve to modulate or change reactivity (Rothbart & Bates, 2006). Rothbart s (Rothbart & Bates, 2006; Rothbart & Derryberry, 1981) temperament model overlaps considerably with other models for studying temperament, but there are several advantages to Rothbart s approach that are important for the present studies. First, Rothbart argues that there should be a clear distinction between positive and negative emotions (Goldsmith et al., 1987; Rothbart & Derryberry, 1981). This distinction is important because approach behaviors tend to be associated with high levels of positive emotion and low levels of negative emotion, whereas withdrawal behavior is associated with high levels of negative emotion (Kagan, Snidman, & Arcus, 1998). The present study will examine positive and negative emotions separately in response to novelty. Second, Rothbart differentiates types of negative emotions, such as distress to limitations (anger reactivity) and distress to novelty (fear reactivity), from one another instead of using broad terms such as fussy. Rothbart, Derryberry, and Hershey (2000) found that high levels of anger during infancy predicted later approach behavior, whereas high levels of fear predicted low approach behavior. Since fear and anger are differentially correlated with approach and withdrawal behaviors, it is particularly important to distinguish between the two emotions. Finally, Rothbart acknowledges that different aspects of temperament come online at different points during development. For example, temperamental approach is present in early infancy, whereas temperamental

30 18 inhibition emerges in the second half of the first year of life (Rothbart, 1988). Each of these fine-grained distinctions in Rothbart s model help lay the foundation for conceptualizing temperament in the present investigation. Approach/withdrawal processes The aspect of temperament most relevant to the present studies is approach/withdrawal. Approach/withdrawal processes underlie individual differences in response to novel or unfamiliar stimuli, such as new toys, new people, new situations, and potentially new foods. Individuals predisposed to respond with approach behaviors are likely to show positive affect and physical movement toward novel stimuli. Conversely, individuals predisposed to respond with withdrawal behaviors are likely to show negative affect and rejection or movement away from the same stimuli (Fox & Henderson, 1999). Work with animals indicates that approach/withdrawal processes are present in all species and serve an adaptive function (Schneirla, 1965). Research with humans reveals that these processes are relatively stable over time (Kagan & Fox, 2006). Approach/withdrawal processes were originally conceptualized as being part of one continuum with the tendency to approach on one end and the tendency to withdraw on the other (Thomas & Chess, 1977). However, neurological theories (Gray, 1982; Gray & McNaughton, 2000) and empirical work on the development of these systems (Putnam & Stifter, 2002; Rothbart, 1988) suggests that approach and withdrawal are controlled by two separate systems. In Gray s (1982) original model, the Behavioral Activation System (BAS) was responsible for organizing approach behaviors in response to reward or

31 19 nonpunishment stimuli, whereas the Behavioral Inhibition System (BIS) was responsible for inhibiting approach to conditioned aversive stimuli that signaled novelty, punishment, and nonreward (Gray, 1991). In a recent revision of the theory (Gray & McNaughton, 2000), the BAS remains largely the same, but the fight-flight-freeze (FFFS) system is now responsible for controlling avoidance and escape behaviors in response to all aversive stimuli. The BIS no longer coordinates an individual s responses to aversive stimuli, but it resolves goal conflicts, often between the BAS and the FFFS (Corr, 2004). Regardless of whether the BIS or the FFFS controls avoidance behaviors, the implications of both versions of the theory are the same for temperament researchers: approach and withdrawal responses are controlled by two separate but interacting systems. Empirical studies investigating the development of approach and inhibition, or the withdrawal behavior of hesitating in response to novelty, suggest that two separate systems underlie these behaviors (Putnam & Stifter, 2002; Rothbart, 1988; Schaffer, Greenwood, & Parry, 1972). Longitudinal studies following infants across the first year of life have revealed that younger infants show indiscriminate approach behaviors toward novel stimuli; they tend to reach toward stimuli at the same speed regardless of whether the stimuli are familiar or novel, or of low or high intensity. However, in the third quarter of the first year, infants develop the ability to inhibit their approach behaviors, and they hesitate before reaching toward stimuli that are unfamiliar or highly intense (Putnam & Stifter, 2002; Rothbart, 1988; Schaffer et al., 1972). The results of these studies demonstrate that approach and inhibition must be controlled by two separate systems that develop at different times during the first year of life (Putnam & Stifter, 2002; Rothbart,

32 ). Further, the development of inhibited behavior also parallels the emergence of fear which is the emotion that motivates individuals to avoid danger (Saarni, Campos, Camras, & Witherington, 2006). Behavioral inhibition One methodological approach to studying approach/withdrawal is the study of behavioral inhibition led by Kagan and colleagues (Garcia-Coll, Kagan, & Reznick, 1984; Kagan et al., 1984; Kagan, Reznick, Snidman, Gibbons, & Johnson, 1988; Kagan & Snidman, 1991). In this approach, children are separated into categories based on their initial responses to unfamiliarity. Inhibited children are characterized by high levels of wariness and distress in response to novel stimuli (i.e. new toys, new people), whereas uninhibited children are characterized by low levels of distress in response to novel objects and high levels of sociability in response to new people. The two classifications include only the most extremely (top 10%) inhibited and uninhibited children (Kagan & Fox, 2006). Kagan and colleagues have demonstrated that differences in behavior between inhibited and uninhibited children are produced by distinctly different neural profiles (Fox & Henderson, 1999; Kagan & Fox, 2006). In particular, it is assumed that the two groups have differential amygdala activation in response to sensory stimuli. The amygdala mediates fear responses by processing input from sensory regions and directing fight or flight behaviors and reactivity (i.e. crying, motor activity). Thus, Kagan believes that differences in behavior in response to sensory stimuli (i.e. visual, auditory, olfactory)

33 21 during infancy may indicate variations in amygdala excitability (Kagan & Fox, 2006). In support of this argument, Kagan has shown that infants who are likely to be later classified as behaviorally inhibited exhibit high levels of motor activity and negative affect in response to sensory stimuli. Conversely, infants who are likely to be later classified as uninhibited show low levels of motor activity and no negative affect in response to the same stimuli (Kagan et al., 1998). Kagan has also shown that his classifications of high reactive and low reactive infants and inhibited and uninhibited children are moderately stable overtime. Longitudinal studies have demonstrated that about one-quarter of high reactive infants are classified as inhibited during early childhood and about one-quarter of low reactive infants are classified as uninhibited during early childhood (Kagan & Snidman, 1991; Kagan et al., 1998). Further, children who were previously classified as behaviorally inhibited at 21 months of age were more likely to be socially inhibited with an unfamiliar peer and more cautious during lab tasks at 4 years of age compared to uninhibited children (Kagan et al., 1984). An additional follow-up study of the same children at 7 ½ years of age indicated that the majority of inhibited children were shy, timid, and quiet, whereas the majority of uninhibited children were sociable with unfamiliar peers and showed greater approach behaviors in novel situations (Kagan et al, 1988). Although Kagan s body of work is the most concentrated effort towards understanding behavioral responses to novelty (Putnam & Stifter, 2005), his approach has several limitations. One such limitation is that his method only includes individuals who exhibit extremely inhibited or uninhibited behavior. This inclusion criteria excludes a substantial portion of his study participants who are not included in analyses. Therefore,