<%BANNER%>

Flavor Learning in Humans through Associations with an Energy-Dense Food Matrix

MISSING IMAGE

Material Information

Title:
Flavor Learning in Humans through Associations with an Energy-Dense Food Matrix
Physical Description:
1 online resource (101 p.)
Language:
english
Creator:
Gardner, Elizabeth A
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Food Science and Human Nutrition
Committee Chair:
Sims, Charles A
Committee Members:
Sommerfield, Linda May
Clark, David G

Subjects

Subjects / Keywords:
acceptability -- flavor -- food -- hglms -- learning -- sensory
Food Science and Human Nutrition -- Dissertations, Academic -- UF
Genre:
Food Science and Human Nutrition thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Humans make food choices based on pleasure and emotion through olfaction. Innate responses to basic tastes and the ability to learn dictate food preference. However, evidence of conditioning a flavor in humans is unclear. The main objective of this study was to investigate mechanisms responsible for positive flavor learning. Participants rated five singularly-flavored 5% sugar solutions and one of two singularly-flavored ice cream samples in two separate studies. Participants from each study were assigned either to guava-flavored ice cream or lychee-flavored ice cream. Participants provided hedonic ratings of all flavored solutions before and after ice cream exposure. Study one consisted of one exposure to the ice cream, while Study two consisted of three exposures. A two-way Analysis of Variance was performed to compare the pre- and post-ratings of the lychee- or guava-flavored solution for the lychee and guava participants, respectively. We hoped to see significant increases in ratings of the lychee solution for lychee participants and in ratings of the guava solution for guava participants. Results show increased hedonic ratings for both lychee-and guava-flavored solutions after ice cream exposure, though no significant increases (a=0.05) were observed. For Study one a near-significant increase in ratings for the guava solution of the guava group was observed (p=0.0548). Results were consistent between both studies. These studies suggest repeated ice cream exposures may be necessary. Additionally, hunger and food matrix may be driving factors for flavor liking. Food choice can be better understood with a broader knowledge of flavor learning and positive reinforcement.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Elizabeth A Gardner.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Sims, Charles A.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045698:00001

MISSING IMAGE

Material Information

Title:
Flavor Learning in Humans through Associations with an Energy-Dense Food Matrix
Physical Description:
1 online resource (101 p.)
Language:
english
Creator:
Gardner, Elizabeth A
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Food Science and Human Nutrition
Committee Chair:
Sims, Charles A
Committee Members:
Sommerfield, Linda May
Clark, David G

Subjects

Subjects / Keywords:
acceptability -- flavor -- food -- hglms -- learning -- sensory
Food Science and Human Nutrition -- Dissertations, Academic -- UF
Genre:
Food Science and Human Nutrition thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Humans make food choices based on pleasure and emotion through olfaction. Innate responses to basic tastes and the ability to learn dictate food preference. However, evidence of conditioning a flavor in humans is unclear. The main objective of this study was to investigate mechanisms responsible for positive flavor learning. Participants rated five singularly-flavored 5% sugar solutions and one of two singularly-flavored ice cream samples in two separate studies. Participants from each study were assigned either to guava-flavored ice cream or lychee-flavored ice cream. Participants provided hedonic ratings of all flavored solutions before and after ice cream exposure. Study one consisted of one exposure to the ice cream, while Study two consisted of three exposures. A two-way Analysis of Variance was performed to compare the pre- and post-ratings of the lychee- or guava-flavored solution for the lychee and guava participants, respectively. We hoped to see significant increases in ratings of the lychee solution for lychee participants and in ratings of the guava solution for guava participants. Results show increased hedonic ratings for both lychee-and guava-flavored solutions after ice cream exposure, though no significant increases (a=0.05) were observed. For Study one a near-significant increase in ratings for the guava solution of the guava group was observed (p=0.0548). Results were consistent between both studies. These studies suggest repeated ice cream exposures may be necessary. Additionally, hunger and food matrix may be driving factors for flavor liking. Food choice can be better understood with a broader knowledge of flavor learning and positive reinforcement.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Elizabeth A Gardner.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Sims, Charles A.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045698:00001


This item has the following downloads:


Full Text

PAGE 1

1 FLAVOR LEARNIN G IN HUMANS THROUGH ASSOCIATION WITH AN ENERGY DENSE FOOD MATRIX By ELIZABETH ANN GARDNER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Elizabeth Ann Gardner

PAGE 3

3 To my loving family and friends

PAGE 4

4 ACKNOWLEDGMENTS I thank my supervisory committee: Dr. Charles Sims, Dr. Linda Bartoshuk, and Dr. Dave Clark. Thank you for guiding me through this study by providing me with your constant knowledge and support. I would especially like to thank Dr. Charles Sims for serving as my major advisor and giving me the opportunity to pursue my graduate degree. I also thank Dr. Bartoshuk for providing me with the extensive support for a project outside the normal realm of food science. Your support and help is greatly appreciated. I thank all my labmates past and present Brittany Martin, Betty Natasha Coello, Jamila Lepore, Eric Dreyer, Dr. Asli Odabasi, Rachel Glintz, and Kelly Brown. Thanks so much with all my informal tasting sessions, especially the preliminary flavor formulati ons. I would not have been able to come up with the sugar solutions without your help. I thank Brittany and Rachel for helping set up my panels. Thank you Eric for helping format my ballot, answering my many questions, helping make (and taste) the ice crea m samples, and handling all the CompuSense issues for the taste panels. Thank you to all for your help and support while I was writing my thesis! I also thank the Sensory Lab crew for setting up and running panels for my studies. Thank you to those who par ticipated and encouraged me during my research. I thank all my panelists for their participation and eagerness to be involved. I thank Dr. Wade Yang and Akshay Anugu for allowing me to use the ice cream batch freezer, answering questions, and offering help I would like to offer a huge thank you to Edward Coryn of Dairy Mix, Inc., St. Petersburg, Florida and Michael Manfredi of Sweet Dreams Ice Cream, Gainesville, Florida for supplying me with ice cream mix and answering all my ice cream questions. Without you this project would not have been a success.

PAGE 5

5 I thank the people in my life who have in some way encouraged me to reach the point I am currently. I especially thank Lorenzo Puentes for motivating me to further my education and encouraging me to complete my thesis. Thank you for your endless support, advice, and listening ear. I thank all my school teachers who helped me get this far. I would especially like to acknowledge Mr. Paul Parker for inspiring me to reach my goals and supplying me with the tools t o write effectively. It is to you I owe my thanks for motivating me to attend the University of Florida and you who helped me write more eloquently. Go Gators! I thank all my friends who have in some way made the experience of working on my graduate degree and thesis easier. I especially thank Perry Fowlkes, Michael Dees, Joe Kirkham, Amanda Clothier, Jordan Ball, Emily Kirkham, Erik Brown, Rachel Hoffman, and Steve Farlekas. Our times together thank Robert stbye for sticking by my side throughout my time at the University of Florida, especially during my graduate career. Your endless support has helped me achieve the most important academic achievement in my life. This process would otherwise have been very difficult and would have seemed impossible. You have been my crutch and my shoulder to lean on. Thank you so much and jeg elsker deg! Finally, I thank my loving family, especially my parents and brothers. Your support and positive example t hroughout my entire life have carried through the years How funny is that?

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBR EVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 2 LITERATURE REVIEW ................................ ................................ .......................... 18 Flavor Perception ................................ ................................ ................................ .... 18 Taste Perception ................................ ................................ ................................ ..... 18 Orthonasal and Retronasal Olfaction ................................ ................................ ...... 19 Taste and Odor Mixtures ................................ ................................ ........................ 20 Hedonics of Taste and Olfaction ................................ ................................ ............. 22 Acquisition of Liking ................................ ................................ ................................ 24 Learning and Preference ................................ ................................ .................. 24 Associative Learning ................................ ................................ ........................ 25 Methods of Sensory Evaluation ................................ ................................ .............. 28 Category Scales ................................ ................................ ............................... 29 Line Marking Scales ................................ ................................ ......................... 30 Direct Scaling Methods ................................ ................................ ..................... 30 Labele d Scales ................................ ................................ ................................ 31 3 RESEARCH METHODS AND MATERIALS ................................ ........................... 34 Study One ................................ ................................ ................................ ............... 34 Study Two ................................ ................................ ................................ ............... 35 Panelist Recruitme nt ................................ ................................ ............................... 36 Food Selection ................................ ................................ ................................ ........ 36 Sensory Laboratory ................................ ................................ ................................ 37 Questionnaire Design ................................ ................................ ............................. 38 Statistical Analysis ................................ ................................ ................................ .. 39 4 RESULT S AND DISCUSSION ................................ ................................ ............... 41 Study One ................................ ................................ ................................ ............... 43 Study Two ................................ ................................ ................................ ............... 48

PAGE 7

7 5 CONCLUSION ................................ ................................ ................................ ........ 76 APPENDIX A QUESTIONNAIRE ................................ ................................ ................................ .. 79 B HGLMS ................................ ................................ ................................ ................... 80 C IRB FORM ................................ ................................ ................................ .............. 81 D COMPUSENSE TEST BALLOT ................................ ................................ ............. 82 LIST OF REFERENCES ................................ ................................ ............................... 97 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 101

PAGE 8

8 LIST OF TABLES Table page 3 1 Demographic information from all panelists ................................ ........................ 40 4 1 Study one guava group mean ratings and p values for flavored solutions ......... 55 4 2 Study one lychee group mean ratings and p values for flavored solutions ......... 55 4 3 Group mean ratings of ice cream samples for study one and study two ............ 56 4 4 Study one guava group mean solution ratings and p values for normal weight group ................................ ................................ ................................ .................. 58 4 5 Study one guava group mean solution ratings and p values for overweight group ................................ ................................ ................................ .................. 58 4 6 Study one lychee group mean solution ratings and p values for normal weight group ................................ ................................ ................................ ....... 58 4 7 Study one lychee group mean solution ratings and p values for overweight group ................................ ................................ ................................ .................. 58 4 8 Overall mean ice cream ratings for normal weight and overweight BMI groups for study one and s tudy two, including p values for comparison of normal weight to overweight for each exposure ................................ ................. 61 4 9 Study one guava group mean solution ratings and p values for female group ... 62 4 10 Study one guava group mean solution ratings and p values for male group ...... 63 4 11 Study one lychee group mean solution ratings and p values for female group .. 64 4 12 Study one lychee group mean solution ratings and p values for male group ..... 65 4 13 Study two guava group mean ratings and p values for flavored solutions .......... 66 4 14 Study two lyc hee group mean ratings and p values for flavored solutions ......... 66 4 15 Study two fruit consumption history ................................ ................................ .... 68 4 16 Study two guava group mean solution ratings and p values for normal weight group ................................ ................................ ................................ .................. 70 4 17 Study two guava group mean solution ratings and p values for overweight group ................................ ................................ ................................ .................. 70

PAGE 9

9 4 18 Study two lychee group mean solution ratings and p values for normal weight group ................................ ................................ ................................ .................. 70 4 19 Study two lychee group mean solution ratings and p values for overweight group ................................ ................................ ................................ .................. 70 4 20 Study two guava group mean solution ratings and p values for female group ... 73 4 21 Study two guava group mean solution ratings and p values for male group ...... 73 4 22 Study two lychee group mean solution ratings and p values for female group ... 73 4 23 Study two lychee group mean solution ratings and p values for male group ...... 73

PAGE 10

10 LIST OF FIGURES Figure page 4 1 Study one guava group mean ratings for flavored solutions with p values indicated above each sample ................................ ................................ ............. 55 4 2 Study one lychee group mean ratings for flavored solutions with p values indicated above each sample ................................ ................................ ............. 56 4 3 Study one distribution of flavored solution ratings for guava group .................... 57 4 4 Study one distribution of flavored solution ratings for lychee g roup .................... 57 4 5 Study one guava group mean solution ratings for normal weight group with p values indicated above each sample ................................ ................................ .. 59 4 6 Study one guava group mean solution ratings for overweight group with p values indicated above each sample ................................ ................................ .. 59 4 7 Study one lychee group mean solution ratings for normal weight group with p values indicated above each sample ................................ ............................... 60 4 8 Study one lychee group mean solution ratings for overweight group with p values indicated above each sample ................................ ................................ .. 60 4 9 Study one guava group mean solution ratings for female group with p values indicated above each sample ................................ ................................ ............. 62 4 10 Study one guava group mean solution ratings for male group with p values indicated above each sample ................................ ................................ ............. 63 4 11 Study one lychee group mean solution ratings for female group with p values indicated above each sample ................................ ................................ ............. 64 4 12 Study one lychee group mean solution ratings for male group with p values indicated above each sample ................................ ................................ ............. 65 4 13 Study two guava group mean ratings for flavored solutions with p values indicated above each sample ................................ ................................ ............. 66 4 14 Study two lychee group mean ratings for flavored solutions with p values indicated above each sample ................................ ................................ ............. 67 4 15 Study two average ice cream ratings for lychee and guava across three sessions ................................ ................................ ................................ ............. 67 4 16 Study two distribution of flavored solution ratings for guava group ..................... 69

PAGE 11

11 4 17 Study two distribution of flavored solution ratings for lychee group .................... 69 4 18 Study two guava group mean solution ratings f or normal weight group with p values indicated above each sample ................................ ................................ .. 71 4 19 Study two guava group mean solution ratings for overweight group with p values indicated above each sample ................................ ................................ .. 71 4 20 Study two lychee group mean solution ratings for normal weight group with p values indicated above each sample ................................ ................................ .. 72 4 21 Study two lychee group mean solution ratings for overweight group with p values indicated above each sample ................................ ................................ .. 72 4 22 Study two guava group mean solution ratings for female group with p values indicated above each sample ................................ ................................ ............. 74 4 23 Study two guava group mean solution ratings for male group with p values indicated above each sample ................................ ................................ ............. 74 4 24 Study two lychee group mean solution ratings for female group with p values indicated above each sample ................................ ................................ ............. 75 4 25 Study two lychee group mean solution ratings for male group with p values indicated above each sample ................................ ................................ ............. 75

PAGE 12

12 LIST OF ABBREVIATIONS ANOVA Analysis of Variance BMI Body Mass Index CTA Conditioned Taste Aversion EC Evaluative Conditioning FNL Flavor Nutrient Learning FSHN Food Science and Human Nutrition g LMS General Labeled Magnitude Scale H edonic g LMS Hedonic General Labeled Magnitude Scale IRB Institutional Review Board LAM Labeled Affective Magnitude LMS Labeled Magnitude Scale ME Magnitude Estimation SAS Statistical Analysis System VAS Visual Analogue Scale

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FLAVOR LEARNING IN HUMANS THROUGH ASSOCIATION WITH AN ENERGY DENSE FOOD MATRIX By Elizabeth Ann Gardner August 2013 Chair: Charles A. Sims Major: Food Science and Human Nutrition Humans make food choices based on pleasure and emotion through olfaction Innate responses to basic tastes and the ability to learn dictate food preference. However, evidence of conditioning a flavor in human s is unclear. The main objective of this study was to investigate mechanisms responsible for positive flavor learning. P articipants rate d five singularly flavored 5% sugar solutions and one of two singularly flavor ed ice cream samples in two separate studies P articipants from each study were assigned either to guava flavored ice cream or lychee flavored ice cream. Participants provided hedonic ratings of all flavored solutions before and after ice cream exposure S tudy one consisted of one exposure to the ice cream while Study two consisted of three exposures. A two way Analysis of Variance was performed to compare the pre and post ratings of the lychee or guava flavored solution for the lychee and guava particip ants, r espectively. We hoped to see significant increase s in ratings of the lychee solution for lychee participants and in ratings of the guava solution for guava participants. Results show increase d hedonic rating s for both lychee and guava flavored solu tions after ice cream exposure

PAGE 14

14 observed. F or Study one a near significant increase in ratings for the guava solution of the guava group was observed (p=0.0548). Results were consistent between both studies Th ese studies suggest repeated ice cream exposures may be necessary. Additionally, hunger and food matrix may be driving factor s for flavor liking Food choice can be better underst ood with a broader knowledge of flavor learning and positive reinforcement.

PAGE 15

15 CHAPTER 1 INTRODUCTION Consumption of reliable, safe, and energy rich foods generates either positive or negative food responses From birth, humans respond positively or negatively to the basic tastes due to genetic predispositions (Shepherd 2006; Adler and others 2000; Yeomans 2008; Drewnowski 1997; Sclafani 1997; Birch 1999; Myers and Sclafani 2006) Taste describes the pure gustatory properties sweet, salty, sour, and bitter (Rozin 1982; Adler and others 2000; Shepherd 2006) Sweet is preferred while bitter and sour are avoided. Innate prefe rences such as these can change throughout life due to learning. For instance, humans learn to like the bitter taste from caffeine and alcoholic beverages. This plasticity in food preference is dictated by the consequences of consuming food. Consequences c an be positive or negative, and repeated associations of food sensory cues with postingestive signals can produce positive or negative learned preferences (Myers and Sclafan i 2006) This production of positive or negative learning preferences through the consequential consumption of food is referred to as conditioning. The ability to condition a positive preference in humans is of particular interest for understanding lear ned food preferences. Although extensively studied in animal subjects, there are limited studies of positive conditioning of food preferences in human subjects In this study, two flavor systems were evaluated for overall liking before and after exposure to an ice cream sample flavored with one of the two flavor systems. Ratings of the flavor systems before and after ice cream exposure were compared to determine if an increase in overall li king occurr ed The ability of humans to learn food preferences stems from making food choices based on pleasure and emotion. The overall sensory experience involves the

PAGE 16

16 combination of taste, touch, smell, vision, and sound. Retronasal olfaction results from the detec tion of odor molecules released from food. During mastication, a food matrix is broken down resulting in the subsequent release of odor molecules (Shepherd 2005) The release of these vol atiles in the back of the mouth and throat results in retronasal olfaction. Flavor results from the combination of taste and retronasal olfaction. T he functions of the human olfactory system therefore dictate what is considered pleasurable. Additionally, f ood choices and preferences are influenced by i nnate responses to basic tastes and our ability to learn. It has been hypothesized that the overall liking for a single flavor can be increased in human subjects through the pairing of that flavor with an inn ately positive reinforcing food matrix (eg. sweet and energy dense). Additionally, flavors not paired with a positive reinforcing food matrix should remain unchanged in level of overall liking. It is hoped that sensory scientists and taste psychophysicist s can benefit from the results of this study. Human food choice and preference can be better understood with a broader knowledge of the mechanisms of flavor learning and positive reinforcement during food consumption Further sensory related research is ne cessary to understand whether negative conditioning for food preferences occur s similarly to positive conditioning. Retronasal olfaction is influential on the enjoyment of food and the pleasure derived from it (Rozin 1982) Human food choices are based on pleasure and emotion and the positive associations created through conditioning are therefore important to dictating preference and overall likability. Influences on food preference may play a role in consumer acceptability. Additionally, the food industr y may benefit

PAGE 17

17 from better understanding our innate responses to basic tastes and the influence of postingestive effects of nutrients in relation to flavors. The main objective of this study wa s to investigate how humans learn to like flavors through positi ve flavor learning. Despite the clear findings of animal studies, evidence of conditioning a flavor preference in humans is less reliable owing to very few studies conducted. Results of this study could support a positive relationship between the pairing o f a flavor with a positive ly perceived food matrix and the i ncreased liking of that flavor, which would provide further understanding of positive food learning and preference in humans

PAGE 18

18 CHAPTER 2 LITERATURE REVIEW Flavor Perception For humans, the consum ption of food is unique. The combination of taste, touch, smell, vision, and sound provide for a complex experience. Flavor is a term used to describe the sensation experienced when consuming a food. As a descriptor in the English language, flavor denotes the combination of mouth and olfactory sensations during food consumption (Rozin 1982) Flavor is influenced by many aspects during mastication, including taste, retronasal olfaction, somatosensation, trigeminal senses, tactile response, and appear ance (Shepherd 2004; Auvray and Spence 2008) Occasionally, additional attributes of a food can influence its perceived flavor. For instance, smell sensed through orthonasal olfaction may influence how a person feels about a flavor. During mastication, a food is broken down and changes in texture occur. This change in texture results in the release of odorants in the mouth which are sensed retronasall y. The surfac es of the mouth sense texture, while t he trigeminal senses detect any temperature changes present. The trigeminal system provides information concerning chemical irritation, temperature, texture, and cons istency (Auvray and Spence 2008) Additionally, the tongue detects the basic tastes sweet, salty, sour, and bitter. Taste Perception Differing from flavor, taste is the perceptual description for only the pure gustatory properties sweet, salty, sour, and bitter (Rozin 1982; Adler and others 2000; Shepherd 2006) Taste combined with the retronasal perception of odor results in the creation of basic flavor perception. It is thought mammals taste many compounds but

PAGE 19

19 only distinguish between several taste modalities, the basic tastes mentioned previou sly. Taste buds are the anatomical structures which contain receptors for mediating our sense of smell. Located throughout the tongue, taste buds contain a small opening called a taste pore. From the taste pore protrude microvilli, which arise from individ ual taste cells (Nelson 1998) Each taste bud contai ns a range of taste receptor cells, which aid in the detection of taste molecules. Taste provides essential information for sensory evaluation of foods, most importantly for the responses evoked during consumption. Taste stimuli elicit responses indicative of aversion (bitter) or attraction (sweet), which lead to pleasure (Drewnowski 1997; Sclafani 1997; Birch 1999; Adler and others 2000; Myers and Sclafani 2006; Shepherd 2006; Yeomans 2008) The essentiality of taste perception is apparent owing it is hardwired in the brainstem from birth. Shepherd (2006) describes the taste pathway as occurring in the brainst em, hypothalamus, taste area of the somatosensory thalamus, and the primary taste cortex. Others refer to taste as a minor sense when compared to other senses, such as smell, due to the channel of only a limited number of sensations (Auvray and Spence 2008) Orthonasal and Retronasal Olfaction The way in which humans perceive odors is through the olfactory system. Olfaction is a dual sense, functioning to perceive the outsi de world as well as objects inside the mouth (Rozin 1982) Orthonasal and retronasal olfaction play prim ary roles in sensing odorants. Orthonasal olfaction provides a direct means of access to an odorant, whereas retronasal olfaction depends on the release of odors from foods during mastication (Shepherd 2005) Odor molecules are the functional pieces of information for olfaction, determining what is sensed as smell or flavor (Shepherd 2005) Odor molecules represent a diverse range of structure for sensory input to the olfaction

PAGE 20

20 system (Mori and others 1999) In mam mals, the main olfactory bulb functions as a processing unit for odor molecules. It contains thousands of glomeruli, which are spherical structures functioning to process signals from synaptic connections excited via odorant receptors (Lancet 1986; Mori and others 1999) The initial event in the perception of an odor involves the interaction of odorous ligands with receptors on sensory neur ons (Shepherd 1994; Vas sar and others 1994; Zhao and others 1998) When humans e xperience an odor, the information is transmitted to the brain by peripheral neurons. The brain processes this information, subsequently developing a representation of that odor, or a template (Vassar and others 1994) Each odor molecule is analyzed by its structural features, which individually activate unique sets of receptors. These receptors activate the appropriate sets of glomeruli and are on olfactory sensory neurons in the olfactory epithelium (Vassar and others 19 94) The olfactory cortex determines which of the numerous glomeruli have been activated. Thus, the brain is able to discern which glomeruli have been activated and puts this into memory as a template. In mammals, glomeruli activation is combinational i n order to respond to a wide variety of odorants. In other words, a single olfactory receptor may recognize multiple odors and an odorant can be recognized by many receptors (Adler and others 2000) Wilson and Stevenson (2003) suggest learning and memory play a vital role in olfactory perception. Olfaction may be experience bound, where exposure alone can affect the extent to which stimuli can be di scriminated. Taste and Odor Mixtures Discrimination and identification of individual components in odor mixtures is thought to be limited to three or four odorants (Laing and Francis 1989; Livermore and

PAGE 21

21 Laing 1996; Jinks and Laing 1999b; Laing and others 2002; Wilson and Stevenson 2003; Marshall and others 2005) Jinks and Laing (1999b) suggest the temporal order of the perception and identity of components cannot be discerned in ternary mixtures. Additionally, it is suggested that the cause may stem from slow processing in olfactory working memory which limits the anal ysis of mixtures to about three components. Several researchers hint memory and learning play significant roles in olfactory perception (Jinks and Laing 1999b; Laing and others 2002; Wilson and Stevenson 2003; Marshall and others 2 005) Rat her than processing odorants and odor mixtures analytically, mammals appear to use synthetic processing. Synthetic processing refers to mammals having a limited ability to identify components within odor mixtures and rather treat mixtures as a whole (Wilson and Stevenson 2003) Original studies on the identification of components within mixtures evaluated two compone nt mixtures. Bartoshuk (1975) studied two, three, and four component mixtures, since the taste mixture interactions outside the laboratory are the most important to understand. Marshall and others (2005) suggests the limited capacity to identify components of mixtures is due to the combination of neural interactions and the limited capacity of working memory. Other research has discussed suppression, or odor masking, as a means for the restricted discrimination and identification of odor mixtures (Bartoshuk 1975; Laing and Francis 1989; Livermore and Laing 1996; Jinks and Laing 1999a; Jinks and Laing 1999b; Laing and others 2002) Suppression encountered in food or odorant mixtures refers to the reduction of perceived intensity in one or more dissimilar odorants. When this occurs, the intensities are suppressed so that the intensity of the mixture is less than the sum of the constituents (Bartoshuk 1975) The occurrence of

PAGE 22

22 suppression may explain an additional mechanism through which the ability to identify mixture components is dampened. It is necessary to distinguish between foods w hich are classified as mixtures and singular flavors. The perception of a singular odorant occurs as described above by orthonasal and retronasal olfaction. A template is created for this odorant as a single unit. However, when humans consume food it is ra re that a singular odorant is being detected alone. Rather, a complex mixture of odorants is consumed at once. Despite being comprised of numerous volatiles or odorants, a flavor as a whole is often thought of and referred to as a singular unit. Referring to a flavor in this way helps distinguish from mixtures, such as a fruit juice blend versus a singular juice. The singular juice is referred to as a singular flavor, despite being comprised of numerous volatiles. Therefore the combination of several flavor s (ie. the juice blend) is commonly referred to as a mixture. Distinction between what is referred to as mixtures and singular flavors aids in the understanding of how humans perceive and learn flavors during consumption. The consumption of these foods dif fers significantly in reference to learning, as described above. The learning of a mixture can be altered when the mixture becomes too complex to distinguish each flavor individually. Thus, if learning of a single flavor is desired, that flavor should be i solated and consumed alone. Hedonics of Taste and Olfaction Unique to some mammals, and particularly humans, is the direct connection between hedonics and food choice or preference. Hedonic qualities of food guides food choice and are therefore a critical determinant of the variety and volume we consume. Consuming food is a source of pleasure, and anticipation of the pleasure alone may be sufficient enough to stimulate it. The initial reaction to the sensory qualities of a food is

PAGE 23

23 to either accept or rejec t that food. Foods are accepted due to some positive feature, causing a learned food preference. The ability of humans to learn has helped shape responses to food, including preferences. Our experiences therefore shape our food choices and can even modify our innate tendency to like sweet and dislike bitter (Yeomans 2008) It is widely accepted that the responses e licited from the four basic taste stimuli sweet, salty, sour, and bitter, are innate in mammals. For example, mammals, including humans, have an innate and universal positive sensory response to sweet tastes (Drewnowski 1997; Sclafani 1997; Birch 1999; Adler and others 2000; My ers and Sclafani 2006; Shepherd 2006; Yeomans 2008) Howe ver, the debate of nature versus nurture in reference to human food preference persists Shepherd (2006) refers to these basic emotions of pleasure from sweet and disgust from bitter not as learned, but instead hard wired from birth. These taste stimuli thus differ from odor images which are learned. These innate responses to the basic tastes are genetic predispositions which influence the development of food preferences (Birch 1999) At birth, sweet taste is preferred while sour and bitter are rejected. Preference for salt is not present at birth, but emerges after several months exhibiting a positive association. These innate responses are demonstrated most easily by infants who pre fer sugar solutions to water and sweeter solutions over less sweet solutions. The pleasure exhibited is measured by observing facial expression reactions to stimuli. For bitter taste, infants show negative facial expressions. This aversion is present at bi rth before learning occurs. Similar to bitter, infants show negative reactions to sour substances. Salt does not reliably elicit a consistent distinctive facial expression or preferential intake which would indicate a

PAGE 24

24 preference or aversion (Birch 1999) The preference for sweet taste is most easily explained by the universal relationship between a sweet taste and foods which are safe, nutritious, and rich in sugar (Yeomans 2008) Likewise, the dislike of bitter tastes reflects the avoidance of poisonous food substan ces. Studies with rats as subjects exhibit similar preference outcomes for sweet taste Ackroff and Sclafani (1991) demonstrated flavor pre ference in rats, proving the food choices of animals are influenced by postingestive effects, or positive reinforcement, of nutrients. Thus, pairing flavors with nutrients which provide higher positive reinforcement may help strengthen the choice of that f lavor. Similarly, studies have found high dietary fat can also condition the preference of flavors. Johnson and others (1991) discovered children acquired flavor preferences based on differences in energy density which were due to differences in fat content. The high energy version of the samples became more preferred after repeated consumption. Fats p rovide food matrices with a characteristic taste and texture. Sensory response to fats first involves the perception of fat soluble volatiles sensed through the nose or mouth (Drewnowski 1997) Oral perception of fats is determined by food texture within the oral cavity during chewing or swallowing. Overall, high fat foods have a desirable sensory response relating to pleasure and preference. Acquisition of Liking Learning and Preference The ability to learn and the innate responses to taste stimuli are the basis of food preference. Survival depends on consuming reliable, safe, and energy rich foods. These foods generate either positive or negative innate preferences. As mentioned, humans have innate responses to the basic tastes. These innate hedonic responses which make up the appetite system likely developed due to evolutionary pressures

PAGE 25

25 (Yeomans 2008) These responses helped humans learn what was safe or harmful. Despite innate responses to basic tastes, there is the potential for modification by exposure. For instance, humans often come to like bitter flavors such as those from alcohol or caffeine. Preference refers to the selection of one item over others. Food preferences are learned by experiences of consuming foods. As mentioned, it is thought humans have genetic predispositions to accept or reject foods based on the basic tastes, thus shaping learned food preferences (Birch 1999) Genetic predispositions for basic tastes place constraints on food preferences, and include developing preferences for foods based on associations with the contexts and consequences of eating the food. In addition to the innate respo nses exhibited towards the basic tastes, innate responses to nutrients appear to exist. Birch (1999) which food preferences reflect innate, unlearned special appetites for needed nutrients, including sugar, salt, fat, protein, carbohydrate, and micronutrients. Most notable are the food preferences for carbohydrates and fat. Ackroff and Sclafani (1991) developed a study to research flavor preferences of rats conditioned by sugars. This study sought to evaluate the reinforcing effects of fruct ose and glucose by examining flavor preference conditioning. Rats pref erred a chow associated with g lucose solutions rather than fructose solutions, suggesting glucose has more potent postingestive reinforcing effects. Johnson and others (1991) demonstrated a preference for fat in a study conducted with samples of differing energy density. Following re peated consumption, subjects acquired flavor preferences for the high energy sample as opposed to the low energy sample Associative Learning Preferences and aversions for food flavors exhibit plasticity throughout an ept sweet and salty and reject bitter and sour.

PAGE 26

26 Differing from innate preferences are acquired food preferences. Learning contributes to the range of individual differences in preferences and their plasticity throughout life. More specifically, humans lear n to associate flavors of foods with the consequences following ingestion. Consequences can be positive and the repeated association of food sensory cues with positive postingestive signals can produce learned preferences. Negative consequences can also oc cur, such as gastrointestinal distress and nausea. Once an aversion is formed, the food will be avoided and learned aversions are not readily abandoned. A learned aversion can result from the single pairing of a food with illness. Learned food preferences form more slowly and are more readily extinguished. These form naturally as a result of normal eating. Learned preferences usually have more subtle effects on food intake than do learned aversions. There are several theories for acquired food preferences, which explain how learning principles may be applied to food preference development. The simplest theory is mere exposure, where flavors increase in preference with repeated exposure without reinforcement (Zellner and others 1983; Myers and Sclafani 2006; Yeomans 2008; Methven and others 2012) Studies examining mere exposure in humans with food stimuli are limited and therefore not fully understoo d. The second theory for acquired food preference is flavor nutrient conditioning or flavor nutrient learning. Also referred to as Pavlovian conditioning, a flavor can become strongly preferred due to a learned association between the flavor and the positi ve consequences of consuming it (Zellner and others 1983; Myers and Sclafani 2006) This type of flavor conditioning can influence preference and increase acceptance. In the same way a flavor can become more accepted and preferred, a flavor can become less accepted through conditioned taste aversion (CTA) CTA results

PAGE 27

27 from the negative consequences of consuming a food, which often result in nausea and gastrointestinal distress. CTA has been studied extensively in rats, proving it is possible to experience aversion due to repeated exposure of a food stimulu s with negative reinforcement. In such studies, rats are predisposed to learn to associate flavors with illness or external visual and auditory cues with external harm. CTA in humans is known anecdotally, most commonly due to illness. Studies creating aver sion in humans are limited due to ethical issues. However, Bernstein (19 78) studied children with cancer who were undergoing chemotherapy. Chemotherapy treatment created nausea and illness, the negative reinforcement necessary to create an aversion. Observed was aversion in children who associated a food with nausea. Positive and negative flavor nutrient learning (FNL) is well understood in animal studies. As mentioned, CTA has been observed in research with human subjects. However, positive flavor nutrient learning in humans has not been extensively studied and is the refore not well understood. A positive flavor nutrient consequence, such as asso ciating a flavor with an energy rich food matrix, would theoretically increase acceptance of that food flavor. Yeomans (2008) suggests several critical features which are necessary for the manifestation of clear likes and dislikes. First, the strongest learning will occur with fo ods and drinks with the largest negative (aversion) or positive (acceptance) effects. Second, learning is progressive, so repeated experiences of flavors with nutrient consequences would be necessary to strengthen the association. Lastly, the degree to whi ch the flavor comes to control behavior depends on the strength of association of flavor and nutrient consequence. Overall, the flavor nutrient learning model predicts repeated exposure will lead to strong liking of a positive flavor

PAGE 28

28 with sources of energy and other positive characteristics. Thus, foods should be a safe source of nutrition and energy. FNL predicts a food which is a major source of energy would strongly reinforce or induce flavor preference. The relationship between energy density and liking for foods supports this prediction (Drewnowski 1997) The final theory for acquired food preference i s the flavor flavor model of evaluative conditioning. This model consists of evaluative conditioning (EC) in which the evaluation of one stimulus changes through association with a second stimulus which is already liked or disliked (Yeomans 2008) Overall, a previously hedonically neutral flavor is associated with a flavor to result in an equivalent change in liking for the previously neutral flavor. Methods of Sensory Evaluation Every day humans attempt to share sensory experiences. Attempts to do so result in the use of intensity descriptors such as weak or strong. Sensory evaluation is a scientific method used to evoke, measure, analyze, and interpret responses to products quantitatively by minimizing biases as much as possible (Meiselman 1993) Numerical data are collected, analyses are performed, and interpretations are made. The dilemma with sharin g sensory e xperiences lies in the great differences in experiences which exist among people. Bartoshuk and others (2004) illustrate this predicament by referring to the most intense pain ever experienced among different groups of people. The woman who has experienced child birth may list this as her most intense pain experienced, while the child with more limited experiences may list a skinned knee or broken bone. Regardless of faults such as this, sensory evaluation attempts to collect the experiences of a group of people into a set of data useful for research and industrial purposes. S ensory evaluation provides quantitative means for measuring human responses. In order to numerically collect data, panelists must be able to quan tify their

PAGE 29

29 responses for the sensory test. There are three classes of sensory evaluation test methods: discrimination, descriptive analysis, and affective analysis (Lawless and Heymann 1999) Discrimination or difference testing answers whether any perceptible difference exists between two types of products. Descriptive analysis quantifies the perceived intensities of the sensory characteristics of a product. Affective testing quantifies the degree of liking or disliking of a product. There are several ways to measure sensory responses, with the most frequent methods including classification, grading, ranking, and scaling (Meilgaard and others 2007) Scaling, the most common sensory evaluation technique in the area of food science, references a scale of numbers for judging samples. Scaling techniques use numbers or words to express the intensity of a perceived attribute, such as sweetness or texture, or reaction to an attribute, such as too soft or just right. When words are used, numbers can be applied so the data can be treated statistically (Meilgaard and others 2007) Scaling involves the application of numbers to quantify s ensory experience, and thus allows the measurement, analysis, and interpretation of responses. Due to the nature of applying numbers to human experience, weaknesses may arise. These weaknesses led to the development of scales by psychologists and psychophy sicists. Category Scales Category scales are a rguably the oldest and most common scaling method and involve the choice of discrete response alternatives to signify increases in sensation, intensity, or degrees of acceptability (Lawless and Heymann 1999) One of the most well known categ ory scales is the Natick nine point scale, which was developed to (Jones and others 1955) V alues given on

PAGE 30

30 category scale s do not necessarily measure the degree to which one sample is greater than another (Meilgaard and others 2007) Category scales therefore do not have ratio properties, meaning 8 is not twice as great as 4. Also problematic is the tendency of panelists to avoid using the endpoints or the extremes of th e scale. Category scales may be arranged as a number or word scale measuring intensity, difference from a reference, or pleasure (hedonic). Category scales often use word descriptors, which can be converted to numbers, making it possible to collect numeric al data to be analyzed. Line Marking Scales Line marking scales involve panelists marking a point on a line to indicate the intensity or amount of some sensory characteristic in relation to anchors. Most commonly, only the endpoints are the labeled anchor s, but there are numerous versions of line marking scales which incorporate additional anchors. The left end usually corresponds with the lower corresponds with upper (Meilgaard and others 2007) Endpoint anc hors are sometimes indented to avoid end effects associated with the reluctance of panelists to use the extremes of the scale (Lawless and Heymann 1999) panelists to indicate the intensity or amount of t he sensory characteristic in relation to some reference sample. Marks made on line scales are converted to numbers by manually measuring the position of each mark using a ruler or computer program. This data can then be analyzed for statistical purposes. D irect Scaling Methods Magnitude estimation (ME), or ratio scaling, involves the unrestricted application of numbers to represent sensation ratios (Stevens 1957; Stevens 1960; Schutz and

PAGE 31

31 Cardello 2001) Humans are able to estimate magnitude differences, but may not be able to easily quantify these differences with a category scale. Scales w ith ratio properties were developed to overcome this problem and were pioneered by S.S. Stevens and colleagues (Stevens 1957; Stevens 1960) Magnitude estimation enta ils the use of positive numbers assigned to a sensation, such that the ratios between the numbers reflect ratios of sensation magnitudes experienced by the panelist. Whereas category and line scales depend on the visual appearance of the sc ale, magnitude e stimation do es not. There are two primary types of magnitude estimation scales. One form involves the use of a reference, in which subsequent sam ples are rated relative to it. The second form does not exploit a provided reference, but instead the panelist ra tes the first sample and ra tes subsequent samples in reference to the first. More often than not, a variation of magnitude estimation scales is most useful in sensory evaluation. One such variation of the magnitude estimation scale is the labeled affecti ve magnitude (LAM) scale. This scale is primarily a magnitude estimation scale, but contains properties unique to nine point hedonic scales. Labeled Scales Labeled scales, like category scales, are often used for their assumed simplicity. A major developme nt in scaling was initiated by Green and colleagues (1993) with the creation of an oral labeled magnitude scale (LMS). The LMS is a la beled scale of sensation intensity based on the concept of a category ratio scale and yields ratio level scaling similar to that produced by magnitude estimation (Green and others 1993; Green and others 1996) It consists of quasi logarithmic spacing and is arranged with verbal anc hors spaced to represent perceptual intervals as determined by ratio scaling.

PAGE 32

32 (Green and others 1996) In a re cent evaluation of the LMS Bartoshuk and others (2005) refer to the elasticity of the LMS which can be used to the advantage of researchers. The elasticity of the LMS is most apparent with the formation of the general labeled magnitude scale (gLMS) in which the all sensations, including pa differentiate between intensities of imagined and experienced sensations, as different meanings may result (Snyder and others 2008) Like the LMS the gLMS is a quasi logarithmic line scale which includes a series of typical intensity descriptors. The versatility and elasticity of the LMS is also apparent from the development of the hedonic g eneral labeled magnitude scale (hedonic gLMS) This scale utilizes properti es of both the LMS and hedonic point scales to achieve effective hedonic evaluation with ratio prope rties. Rather than rating with a nine o utilizes a numbered scale similar to the gLMS. The hedonic gLMS u tilizes the anchor (Bartos huk 2010) The hedonic gLMS is most useful when the comparison of groups is desired (Bartoshuk and others 2005) A final variation of the magnitude estimation scale is the labeled affective magnitude scale (L AM). This scale was developed by Schutz and Cardello (2001) for the specific p urpose of assessing food likability The LAM scale encompasses reliability and sensitivity similar to the nine point hedonic scale, but with minimized endpoint compression similar to magnitude estimation. This scale allows

PAGE 33

33 greater discrimination among products and the ability for conclusions to be made about paneli disliking. Addi tional common scales include the visual analogue scale and magnitude matching. The visual analogue scale (VAS) is a labeled scale with ratio properties which eliminates intermediate descriptors, leaving only the minimum and maximum descriptors (Bartoshuk and others 2003; Bartoshuk and others 2004; Bartoshuk and others 2005) Magnitude matching involves selecting a standard unlikely to be related to taste, and requires panelists to express perceived taste intensities relative to that standard (Bartoshuk and others 2004) This me thod of scaling seeks to solve the problem of making comparisons across individuals or groups. Thus, it may be possible to study the sensory characteristics of foods and link these to behavior.

PAGE 34

34 CHAPTER 3 RESEARCH METHODS AND MATERIALS Study One Ninety seven panelists participated in the first study. Prior to participation, potential participants were screened for fruit consumption history and any food allergies (Appendix A) Panelists were chosen based on having never consumed lychee and/or g uava There were 34 male and 63 female participants ranging in age from 13 to 62. The average age was 24, while most panelists were 18, owing to the fact that most panelists were undergraduate or graduate students. Panelists also identified their ethnic backg round, race, and incidence of ear infection as s hown in Table 3 1. The first study consisted of singularly flavored sugar solutions. Strawberry, peach, raspberry, lychee, and guava flavor systems were added in appropriate concentrations (to provide a natur al flavor) to a 5% sucrose solution. Thus, each panelist was presented with five individual flavored sugar solutions. Two batches of ice cream samples were prepared for panelists. One batch was formulated using the same lychee f lavor system used to prepare the lychee flavored sugar solution The second batch was formulated using the same guava flavor system used to prepare the guava flavored sugar solution. The concentration of flavor added to the ice cream was determined through preliminary informal testin g. Specific information regarding the formulations of the flavored solutions and ice cream samples will be discussed later in this chapter. Panelists were first presented with the five singularly flavored sugar solutions and were asked to rate their level of liking on the Hedonic General Labeled Magnitude Scale (Hedonic gLMS) (Appendix B). In a separate session seven days later 47

PAGE 35

35 panelists were given the lychee flavored ice c ream sample and the remaining 50 were given the guava flavored ice cream sample. Panelists were encouraged to eat the entire portion of ice cream. Panelists were aske d to rate this sample for overall likability using the Hedonic gLMS Two days a fter exposure to the ice cream sample, panelists rated each of the five flavored sugar solut ions again for overall likeability using the Hedonic gLMS. Panelists were assigned separate thirty minute session s to perform each of the three tasting and ratings. Verbal and written instructions were provided. Study Two Study two was completed with 104 new panelis ts. This study was co nducted similarly to study one. However, participants were not screened for fruit consumption prior to participation to avoid narrowing demographic data Potential participants were screened for any food allergies. Table 3 1 inc ludes the demographic data of panelists from this study. There were 42 male and 62 female participants ranging in age from 18 to 62. The average age was 26.5, while most panelists were 21 years of age Panelists were presented with the same food item s in this study. All samples were prepared at concentrations matching those from study one. The five singularly flavored sugar solutions (strawberry, peach, raspberry, lychee, and guava) were rated for overall likeability before ice cream consumption and a fter ice cream consumption using the Hedonic gLMS. Panelists were again assigned to either lychee flavored ice cream or guava flavored ice cream. Fifty two panelists were assigned to lychee flavored ice cream and 52 panelists were assigned to guava flavored ice cream. Panelists were given the appropriate ice cream sample to rate for overall likeability using the Hedonic gLMS on three subsequent days. Thus, p articipants were exposed to the assigned ice cream sample three times instead of one time. Two days after the final ice cream

PAGE 36

36 tasting, panelists rated the flavored sugar solutions again for overall likeability on the Hedonic gLMS Panelists were assigned s eparate thirty minute session s to perform each of the five tasting and ratings. Verbal and written instructions were provided. Panelist Recruitment The University of Florida Health Science Center Institutional Review Board (IRB) approved the study protoco l. Pa rticipants (staff and students) were recruited from the University of Florida. Flyers advertising the study were posted at the sign in area of the Food Science and Human Nutrition (FSHN) Sensory Laboratory to inform potential panelists about the study After gathering contact information from interested panelists, each panelist was screened for consumption history (study one only) and any food allergies. Some panelists were inelig ible to participate in the studies due to the following reasons: failure to pass the consumption history screener (study one) they were lactose intolerant and unable to consume the ice cream sample, had an allergy which prevented their participation, were vegan and refused to consume the ice cream sample, and/or were no longer interested in the study. During the first session of each study panelists reviewed details of the study and signed the IRB consent form (Appendix C ). Panelists were free to withdraw from the study at any time with no consequences. Panelists were compensa ted for participation in each session of the study. Food Selection The fruit flavors u tilized in this study were obtained from Firmenich Flavor and Perfume Company, Plainsboro, New Jersey. Flavors selected included strawberry, peach, raspberry, lychee, an d guava. These flavors were individually added to a sugar solution (5% sucrose weight/volume) to obtain singularly flavored sucrose solutions. Informal preliminary panels were held to determine the appropriate concentration of

PAGE 37

37 eac h flavor in the sugar solu tion. The following concentrations of flavors were added to the 5% sugar solution to create the singularly flavored sucrose solutions (L flavor/100 mL sugar solution) : 500 L/100mL (strawberry); 400 L/100mL (peach); 500 L/100mL (raspberry); 400 L/100mL (guava); 175 L/100mL (lychee). A commercial plain (flavorless) ice cream mix was obtained from Dairy Mix, Inc., St. Petersburg, Florida. The ice cream mix utilized was 16% fat ice cream mix. Per 100g sample of ice cream there were 16.02g total fat, 9.914g saturated fat, 19.16g sugars, and 236.6 kCalories (Dairy Mix, Inc. 2006 ) A Carpigiani LB502 Coldstone Creamery batch freezer was u tilized to produce the flavored ice cream batches. Informal preliminary panels were held to determine the appropriate concentration of flavor added to the ice cream mix prior to freezing. The following concentrations of flavors were added to the plain ice cream mix to create the singularly flavor ed ice cream samples (L flavor/100mL ice cream mix): 300 L/100mL (guava); 150 L/100mL (lychee). Sensory Laboratory Panelists evaluated samples in the University of Florida Sensory Laboratory. Panelists evaluated samples in individual booths which were equipped with computer workstations. Panelist data was collected with a s ensory software program (Compus ense five Sensory Analysis Software for Windows Compusense Inc., Guelph, Canada ). Unsalted crackers and purified water were provided to the panelists to encourage rinsing of the palate before and between samples. Samples were assigned random three digit codes and served in four ounce portion cups. Panelists were given approximately two ounces of each flavored sugar solution and approximately four ounces of ice cream. Panelists were presented with all five flavored

PAGE 38

38 sugar solutions according to Williams Design (Williams 1949) E ach sample was presented i n each presentation position an approximately equal number of times. Questionnaire Design At the beginn ing of the first session, panelists were asked a series of demographic questions. The demographic data collected included gender, age, height (in feet and inches), weight (in pounds ), ethnic background, and occurrence of otitis medi a (ear infection). The p mass index (BMI). BMI (lbs/in 2 ) was calculated using the following equation: ((weight (lbs)/height 2 (in)) x 703) (National Institute of Diabetes and Digestive and Kidney Diseases 2012) The Hedonic gLMS was used in this study (Bartoshuk 2010; Bartoshuk and others 2005) All panelists were trained to use this scale to measure and quantify overall personal hedonic liking of each sample. Panelists were encouraged to write personal scale anchors on the provided paper scale as well as type their anchor descriptors into the com puter program (Appendix D). This scale allows panelists to develop a anchor (0). Paneli sts were able to select any numerical integer on this scale when rating samples. When developing the personalized scale, panelists identified personal clude death of a loved one, serious illness, or failing to achieve a goal. with loved ones, and winning an award. After developing the personalized scale, panelists were asked to practice using their scale by rating a range of differing

PAGE 39

39 experienced During evaluations, panelists used this personalized scale to rate their overall liking of the flavored sugar solutions and ice cream samples. Statistical Analysis The main objective of this study was to determine the differences in hedonic rating of flavored sugar solutions before (pre rating) and after (post rating) exposure to ice cream samples. All flavors were analyzed separately for overall likeability ratings Raw data were collected by the sensory software program CompuSens e. These data were transferred to Microsoft Excel for sorting and pre paration for statistical analyse s. The statistical analyses were performed using S tatistical A nalysis S ystem (SAS) 9.3 (SAS Institute Inc., Cary, NC, USA). A two way Analysis of Var iance (ANOVA) using to determine if there were significant differences at a p value < 0.05 between the pre and post ratings of the flavored sugar solutions. rating s we re compared to means of post ratings to determine if a significant difference occurred. Data were also sorted based on gender (males versus females) and BMI using ANOVA. Two different BMI groups were created, which compared norma l weight, including underwe ight (BMI value less than 25) and overweight (BMI value of 25 or more) (National Institute of Diabetes and Digestive and Kidney Diseases 2012) Additional classification variables including race, ethnic background, and incidence of ear infection were not studied due to insufficient populations.

PAGE 40

40 Table 3 1. Demographic information from all panelists Study 1 Study 2 N = 97 N = 104 Gender Male 34 42 Female 63 62 Ethnic Background Hispanic 12 14 Non Hispanic 85 90 Race White or Caucasian 64 55 Black or African American 11 5 Native American, Alaska Native, Aleutian 0 0 Asian/Pacific Islander 15 40 Other 7 4 Incidence of Ear Infection No 63 84 Yes, but not serious 25 15 Yes, required antibiotics more than once 6 2 Yes, required tubes in ears 3 3

PAGE 41

41 CHAPTER 4 RESULTS AND DISCUSSION There were five treatments in both studies As previously discussed, t reatment types differed in sample flavor. Lychee and guava served as experimental flavors, while strawberry, peach, and raspberry served as control flavors (flavors which were not paired with the ice cream sample) Five singularly flavored sugar solutions were consumed and rated by panelists. Two singularly flavo red ice cream samples were consumed and rated by panelists. Half of the panelists were assigned to a lychee flavored ice cream, while the remaining panelists were assigned to a guava flavored ice cream. Panelists only consumed and rated the ice cream sampl e to which they were assigned. By pairing only the experimental flavors with a high energy food matrix (ice cream), we expected to see the hedonic ratings of only these flavors (lychee and guava) increase after exposure. Since the remaining flavors were no t paired with the ice cream samples, ratings were expected to remain the same after ice cream exposure. Data were analyzed as an overview of all panelists in a given study and were separated into groups based on the flavor of ice cream sample received. Ad ditionally, panelists were further separated into groups based on gender and body mass index ratings of the flavored sugar solutions before and after ice cream exposure wer e compared for a given panelist. Study one was repeated as Study two to determine if the amount of exposure, or volume, of the high energy food matrix affected the overall hedonic ratings of samples. Possible error sources included palate exhaustion, perso nal likeability of the samples, familiarity with the sample flavors, considerable contrast between solutions and assigned ice cream sample, and setting. The remainder of this

PAGE 42

42 chapter will discuss results from each study, focusing on the comparison of the r atings of flavored sugar solutions prior to and following exposure to the assigned ice cream sample. The hypothesis that the hedonic rating of two novel flavors will increase after being paired with an innately positive food matrix (ice cream) was rejected However, changes in overall likeability were observed showing increases in the overall hedonic ratings of the experimental flavors. Although the previous hypothesis was not accepted, the hypothesis was accepted that the hedonic ratings of flavors not paired with the ice cream (strawberry, peach, and raspberry) will not be significantly increased after the ice cream exposure Significan t differences were observed for two separate control flavors (flavors not paired with the ice cream) However, the significant differences observed for these flavors were significant decreases between pre rating and post ratings for the flavored s olution. More specifically, the overall post ratings of these flavors were significantly lower than the pre ratings. The changes in overall hedonic ratings which were observed for the experimental flavors were not as clear ly present with the remaining flav ors. However, the presence of increased ratings between pre and post ratings for experimental flavors is promising for future research, as this demonstrates the initial stages of flavor learning in human subjects. Since a significant increase in the ratin gs of the experimental flavors was not observed it is evident changes to the experimental design are necessary. Panelists were exposed to the assigned ice cream sample once in Study one, while panelists in Study two were exposed to the assigned ice cream s ample three times. The increased exposure to the flavored high energy food matrix did not produce significantly increased

PAGE 43

43 ratings of the experimental flavors. Thus, the number of exposures alone does not significantly contribute to hedonic ratings of flavo rs. It is possible these results can be explained by earlier research indicating hunger as a noteworthy factor for food likeability and preference (Ackroff and Sclafani 1991; Gibson and others 1995) Additionally, strikingly low ratings were obs erved for the flavored sugar solutions while the ratings for the ice cream samples were at an expected range. It is possible this can be explained by the personal likeability of samples among panelists, rather than poor scale usage. The food matrix for the flavors (ie. the flavored sugar solutions) may not have provided the most natural food matrix for consumer likeability testing. Likewise, panelists' familiarity with the experimental flavors may have contributed to the outcome of the results. The actual c hanges in overall hedonic ratings for all samples will be further discussed throughout this chapter. Study One Ninety seven panelists participated in Study one. They consumed and rated all five singularly flavored sugar solutions before and after one expo sure to an assigned ice cream sample. Potential participants were screened for fruit consumption history and only those who had never consumed lychee or guava in any form participated. Data were sorted according to ice cream sample grouping. Panelists who were assigned guava flavored ice cream were sorted separately from panelists who were assigned lychee flavored ice cream. Within the ice cream groupings, the ratings of the flavored solutions were sorted to compare pre ice cream exposure (pre rating) to po st ice cream exposure (post rating) for all five flavors. The mean ratings for overall likeability for the flavored sugar solutions are presented in Table 4 1 for the guava group and Table 4 2 for the lychee group. No

PAGE 44

44 significant differences were observed at p<0.05. However, at p<0.10 a significant increase (p=0.0548) for the guava flavored solution was observed for the guava group. For both the guava group and lychee group, no other significant differences were observed in hedonic ratings for the flavored solutions. However, some variations in overall likeability were observed The following changes in hedonic ratings were observed for the guava group data: decrease between mean pre rating (17.48) and mean post rating (14.04) for strawberry; slight increase between mean pre rating (11.42) and mean post rating (15.72) for peach; slight increase between mean pre rating (2.68) and mean post rat ing (5.14) for raspberry; increase between mean pre rating ( 7.02) and mean post rating ( 1.44) for lychee. These variations in hedonic ratings do not necessarily indicate near significance, as indicated by p values (Table 4 1). All p values for the guava group, with the guava solution as an exception, do not indicate significance or near significance. Figure 4 1 illu strates the previously discussed means and p values. Similar variations in hedonic ratings were observed in the data for the lychee group: decrease between mean pre rating (24.62) and mean post rating (18.89) for strawberry; nearly no difference between me an pre rating (12.66) and mean post rating (12.26) for peach; increase between mean pre rating (5.79) and mean post rating (11.00) for raspberry; increase between mean pre rating ( 4.00) and mean post rating (2.68) for lychee; increase between mean pre rat ing (0.11) and post rating (7.34) for guava. The differences between pre rating and post rating means do not indicate significance or near significance, as indicated by p values (Table 4 2). Figure 4 2 illustrates the means and p values discussed for the l ychee group.

PAGE 45

45 The food samples chosen for this study were designed to present an isolated flavor in the simplest form possible. It was desirable for panelists to be exposed to one flavor per sample. Thus, singularly flavored sugar solutions were created. Fl avors were added to a 5% sucrose solution to create the most natural representation of that flavor possible. Additionally, 5% sucrose was utilized to mask any bitterness or off flavors present in the concentrated flavor systems The high energy food matrix chosen to pair with the experimental flavors was 16% fat plain (unflavored) ice cream mix. The same guava and lychee flavor systems were added separately to create the two individual flavored ice cream samples. The average likeability ratings for each ice cream sample are presented in Table 4 3. These results suggest the ice cream samples were highly liked. Hedonic ratings indicate the two samples were liked nearly to the same degree, with an average lychee rating of 43.83 and average guava rating of 48.90 (Table 4 3). An average rating of nearly 50 would be on the upper end of the scale, as it ranges from 100 to 100. This corresponds closely with scale generation demographic data collected we have concluded the ice cream samples were strongly liked by panelists The singularly flavored sugar solutions, however, were rated considerably lower. The average ratings for the flavored solutions range from 7.02 to 24.62, alt hough the median average rating of all the samples is approximately 6.50. The distribution of how panelists rated the flavored solutions are shown graphically for the guava group and lychee group in Figure 4 3 and Figure 4 4, respectively. This range of av erage ratings for the solutions falls mostly on the positive side of the scale, but is very low. Therefore, it is highly likely most panelists did not enjoy the flavored solutions. Though in theory the

PAGE 46

46 sugar solutions would not need to be initially highly liked, it is apparent from the data that if panelists have initial ratings indicative of disliking, conditioning a stronger liking of that flavor would be very difficult. The same data set was sorted by gender and body mass index (BMI). Associations have b een shown between measures of body mass and reported preference for high fat foods (Rissanen and others 2002) which may relate to higher hedonic ratings. Similar studies have indicated differences in responses to food stimuli between men and women. Women tend to have a more robust response to food related cues than men (Cornier and others 2010) including normal weight (BMI values less than 25) and overweight (BMI values 25 or greater). Sixty four p anelists (66%) were considered normal weight and thirty three (34%) were classified as overweight. Based on previous evidence of an association between higher body mass and increased preference for high fat foods, we expected to see a stronger association of the assigned experimental flavor with the corresponding ice cream sample, and thus a more notable increase in sugar solution ratings for the assigned ice cream flavor for the over weight group. Our results found no associations between BMI values and sig nificance of hedonic ratings for the flavored solutions Table 4 4 and Table 4 5 show the mean flavored solution ratings and p values of the guava group for normal weight and overweight groups respectively. Table 4 6 and Table 4 7 show the mean flavored s olution ratings and p values of the lychee group for normal weight and overweight groups respectively. Additionally, Figures 4 5, 4 6, 4 7, and 4 8 graphically represent the same mean ratings and p values. Additional analyses were

PAGE 47

47 performed to determine i f any differences existed between BMI groups for ice cream hedonic ratings. The overall likeability ratings for the ice cream samples were averaged across the normal weight group and overweight group without separation by ice cream flavor A simple t test was conducted to determine if any significant differences in overall ice cream hedonic ratings occurred between normal weight and overweight BMI groups. Data are presented in Table 4 8, showing the overall mean ice cream ratings for the normal weight group and the overweight group and the p value. Results indicate the mean ice cream rating for the overweight group does not differ significantly (p<0.05) from the normal weight group ice cream ratings. The overall average ice cream hedonic rating for the norma l weight group was 42.47, while the average was 54.15 for the overweight group. There was a near significant difference (p=0.0678) but no significant difference was seen between BMI groups. Since all panelists self reported their weight, it is possible some values were incorrect and varied from their actual weight. Additionally, BMI as an indicator for body type is not sensitive to adiposity levels. It is a reference comparing weight to heigh t and therefore does not necessarily indicate adiposity levels. Although no significant differences were observed, it is important to note variations in hedonic ratings were observed in the data which closely resemble the variations observed in the raw dat a set (no gender or BMI separation). For instance, the normal weight BMI group for the lychee group shows an increase in the mean rating from pre rating (4.22) to post rating (12.11), as shown in Table 4 5 and Figure 4 6. The consistency of the hedonic rat ings is promising for future research in this area.

PAGE 48

48 Separation of data by gender was selected due to differences in food likeability and preferences between men and women. Sixty three females (65%) and thirty four males (35%) participated in this study. We expected to see differences in overall likeability between males and females. More specifically, we hoped to see whether one gender rated experimental flavors significantly higher after exposure to the assigned ice cream sample. Overall, the results do no t show clear differences in ratings among males and females. However, notable changes in hedonic ratings do exist in the data. Most notably, females of the guava ice cream group rated the guava sugar solution significantly higher at p<0.10 on average (p=0. 0911) from pre rating ( 11.18) t o post rating ( 1.55) (Table 4 9 and Figure 4 9). The males of the guava group did not experience as strong a change in ratings although a slight increase occurred from the pre rating (1.71) to t he post rating (8.06) (Table 4 10 and Figure 4 10). Similar variations in hedonic ratings were observed for the female lychee group (Table 4 11 and Figure 4 11) a nd male lychee group (Table 4 12 and Figure 4 12), exhibiting slight increases from pre rati ng to post rating. Though the variability in these ratings do not exhibit significant differences at p<0.05, they indicate differences which co uld eventually show significance in average ratings and thus in overall likeability of samples. Additionally, these variations are consistent w ith the previously discussed data for Study one. The consistency of the changes in hedonic ratings within this data set demonstrates auspicious results which can be applied to future research in this area. Study Two Study two followed the same experimenta l design as Study one, except there were five total sessions, three of which involved ice cream exposure. Additionally, 104 new panelists were recruited since those who had already participated were no longer

PAGE 49

49 eligible. Participant recruitment did not invol ve screening for lychee and guava consumption. However, data were collected for history of fruit consumption for informational purposes. Potential participants were not screened for lychee and guava consumption to avoid narrowing demographic data. Beyond t hese differences, all other methods were kept identical. As in Study one, these panelists consumed and rated five singularly flavored sugar solutions (strawberry, peach, raspberry, lychee, and guava) and were assigned to either lychee flavored ice cream or guava flavored ice cream, after which they rated the sugar solutions again. Panelists were exposed to the assigned ice cream sample three times. The mean ratings for the flavored sugar solut ions are presented in Table 4 13 fo r the guava group and Table 4 14 for the lychee group. No significant increases in average ratings were observed at p<0.05 or p<0.10 for the guava flavored solution. However, significant decreases (p<0.05) were observed for two of the flavored solutions. The peach flavored solution dec reased from the pre rating (8.12) to post rating ( 0.42). Likewise, the lychee flavored sugar solution decreased from the pre rating (6.87) to the post rating ( 0.33). The following changes in overall hedonic ratings were observed for the guava group data: decrease between mean pre rating (3.17) and mean post rating ( 2.00) for strawberry; significant decrease between mean pre rating (8.12) and mean post rating ( 0.42) for peach; slight decrease between mean pre rating (6.04) and mean post rating (3.21) for raspberry; significant decrease between mean pre rating (6.87) and mean post r ating ( 0.33) for lychee; increase between mean pre rating ( 4.15) and mean post rating ( 0.35) for guava. With the exception of peach and lychee, these variations in hedonic ra tings are very small and do not necessarily

PAGE 50

50 indicate near significance, as i ndicated by p values (Table 4 13 ). However, it is imperative to note that the guava flavored solution is the only solution which experienced an overall increase in mean ratings. Al l other solutions decreased in mean ratings between pre and post ratings. Figure 4 13 shows the previously discussed means and p values. Similar variations in hedonic ratings were observed in the data for the lychee group, although some differences were o bserved: slight decrease between mean pre rating (8.23) and mean post rating (6.15) for strawberry; slight increase between mean pre rating (12.33) and mean post rating (14.23) for peach; near significant increase (p=0.0535) between mean pre rating (3.62) and mean post rating (10.60) for raspberry; slight increase between mean pre rating (14.71) and mean post rating (14.94) for lychee; slight decrease between mean pre rating (3.79) and post rating (1.50) for guava. As with the guava group, the differences b etween pre rating and post rating means are very small and therefore do not indicate significance, as i ndicated by p values (Table 4 14 ). However, an increase of near significance was observed for the raspberry flavored solution. This increase could relate to possible error sources, which will be discussed later in this section. Additionally, it is necessary to note that the mean rating for the lychee flavored solution increased between pre and post ratings. Figure 4 14 illustrates the means and p values d iscussed for the lychee group. The food samples tested in this study were identical to those tested in Study one. As indicated previously in the Study one discussion, the samples tested in this experiment were designed to provide a clean and simple system to isolate single flavors. Although this objective was accomplished, the flavored solutions were not

PAGE 51

51 entirely the most natural food system for a consumer taste panel. As with Study one, results for Study two have exhibited the vast differences between flav ored solution ratings and ice cream ratings. Overall, it is apparent the panelists highly liked the ice cream samples, while the flavored solutions were rated considerably low. The average ratings for each ice cream sample across the three separate ice cre am sessions are shown in Table 4 3 and Figure 4 15. The average hedonic ice cream ratings range from 32.89 to 39.64. This range indicates panelists liked the ice cream samples substantially. This corresponds with scale generation demographic data collected for panelists which were rated extremely low, with the average ratings ranging from 4.15 to 14.94, with a median average rating of 4.91. This range is similar to the ra nge observed in Study one. However, the mean ratings between pre rating and post rating for the flavored solutions more frequently decreased in Study two than in Study one. The decreases in mean ratings were more pronounced, often decreasing from a positiv e to a neg ative average rating (Table 4 13 ) which did not occur as frequently in Study one. This could be attributed to the slight variation in experimental design for Study two as compared to Study one. Panelists in Study two were exposed to the assigned ice cream sample three times, whereas pane lists in Study one were exposed only once. We expected the increased exposure in Study two would produce a greater increase in the ratings of the flavored solution which coordinated with the assigned ice cream sample (guava or lychee) due to greater reinfo rcement of the assigned flavor from the ice cream sample. However, this did not occur and the post ratings of nearly all the flavored solutions were lower than the pre ratings. We believe this occurred due to the great contrast between

PAGE 52

52 the solutions and th e assigned ice cream sample. The panelists tasted all five flavored solutions followed by three sessions of tasting flavored ice cream of a considerably high fat content. Following these exposures, panelists again tasted the flavored solutions. It appears from the low overall ratings of the flavored solutions that panelists did not highly like the samples, and in some cases disliked the samples. This low level of likeability and dislike of the samples appears to have been magnified in the panelists mind s a fter being exposed to an ice cream sample which was liked so highly and perceived as very pleasant. Thus, going from tasting ice cream to tasting the solutions may have caused the panelists to give an exaggeratedly low rating. It is evident this occurred w hen considering the results discussed previously for the guava group ratings where the peach and lychee flavored solutions decreased significantly (p<0.05). An additional difference between Study one and Study two was the screening criteria for potential participants. Panelists were not chosen based on fruit consumption history, as in Study one. Instead, anyone could participate, regardless of their previous exposure to lychee or guava. Table 4 1 5 shows the fruit consumption history data collected for Stud y two. Nearly 80% of panelists had previously consumed lychee or guava, which may have had some effect on the outcome of the data. Although the concentrated flavor systems used in this study are unique, familiarity with the flavor may have affected the pan Study two, although the need for panelists who have never been exposed to a particular flavor may not have been necessary. Panelists in Study one had no exposure to lychee or guava and the results are very similar to those of Study two. However, the slight differences in the ratings, particularly the decrease in ratings between pre ratings and

PAGE 53

53 post ychee flavors. Thus, this may indicate the need for flavors to be novel to panelists. We expect altering previously formed opinions about a food may be more difficult than altering the opinion about novel flavors. Figure 4 16 and Figure 4 17 show the distr ibution of how panelists rated the flavored solutions for the guava group and lychee group, respectively. This range of average ratings for the solutions falls mostly on the positive side of the scale, which is similar to distributions observed in Study on e. The same data set from Study two was also sorted by gender and BMI. As with and overweight (BMI values 25 or greater). Seventy eight panelists (75%) were considered n ormal weight and twenty six (25%) were classified as overweight. Our results found no associations between BMI values and significance of hedonic ratings. Table 4 1 6 and Table 4 17 show the mean flavored solution ratings and p values of the guava group for normal weight and overweight groups respectively. Table 4 18 and 4 1 9 show the mean flavored solution ratings and p values of the lychee group for normal weight and overweight groups respectively. Additionally, Figure s 4 18 4 19 4 20 and 4 21 graphi cally represent the same mean ratings and p values. Similar results were observed as were found in Study one. As with Study one, analyses were performed to determine if BMI groups differed in overall hedonic ice cream ratings (no separation of guava from l ychee flavors). Table 4 8 shows the average overall ice cream hedonic rating for each exposure and the p value corresponding to the comparison of the normal weight group to the overweight group. No significant differences were observed for any of the three exposures when hedonic ratings of the ice cream for the normal weight

PAGE 54

54 group were compared to those of the overweight group Although no significant differences were observed, it is important to note some variations in hedonic ratings were observed in the data which closely resemble th ose observed in the raw data set (no gender or BMI separation) for Study two. Separation of data by gender was selected again in Study two to discover if any differences existed between males and females. Sixty two females (60 %) and forty two males (40%) participated in this study. No significant differences in ratings among males and females were observed. Table 4 20 and Table 4 21 show the mean flavored solution ratings and p values of the guava group for female and male grou ps respectively. Table 4 2 2, and 4 23 show the mean flavored s olution ratings and p values of the lychee group for female and male groups respectively. Additionally, Figure s 4 22 4 23 4 24 and 4 25 graphically represent the same mean ratings and p val ues. The changes present in hedonic ratings are very small and do not indicate near significance. However, these results are consistent with the previousl y discussed data for Study two.

PAGE 55

55 Table 4 1. Study one guava group mean ratings and p values for flavored solutions Strawberry Peach Raspberry Lychee Guava Pre rating 17.48 11.42 2.68 7.02 6.80 Post rating 14.04 15.72 5.14 1.44 1.72 P value 0.4696 0.2062 0.6428 0.2207 0.0548 Table 4 2. Study one lychee group mean ratings and p values for flavored solutions Strawberry Peach Raspberry Lychee Guava Pre rating 24.62 12.66 5.79 4.00 0.11 Post rating 18.89 12.26 11.00 2.68 7.34 P value 0.1334 0.9314 0.2159 0.242 0 0.1943 Figure 4 1. Study one guava group mean ratings for flavored solutions with p values indicated above each sample -10 -5 0 5 10 15 20 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Average Ratings for Flavored Solutions: Guava Group Pre-Rating Post-Rating 0.4696 0.2062 0.6428 0.0548 0.2207

PAGE 56

56 Figure 4 2. Study one lychee group mean ratings for flavored solutions with p values indicated above ea ch sample Table 4 3. Group mean ratings of ice cream samples for study one and study two Study 1 Study 2 Guava Lychee Guava Lychee N = 50 N = 47 N = 52 N = 52 Exposure 1 48.90 43.83 32.89 39.64 Exposure 2 N/A N/A 33.15 37.17 Exposure 3 N/A N/A 36.23 37.52 -10 -5 0 5 10 15 20 25 30 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Average Ratings for Flavored Solutions: Lychee Group Pre-Rating Post-Rating 0.1334 0.9314 0.2159 0.1943 0.2420

PAGE 57

57 Figure 4 3. Study one distribution of flavored solution ratings for guava group Figure 4 4. Study one distribution of flavored solution ratings for lychee group 0 2 4 6 8 10 12 14 -100 -50 0 50 100 Rating Frequency Individual Raw Rating Distribution of Guava Solution Pre and Post Ratings: Guava Group Pre-Rating Post-Rating 0 1 2 3 4 5 6 7 8 9 10 -100 -50 0 50 100 Rating Frequency Individual Raw Rating Distribution of Lychee Solution Pre and Post Ratings: Lychee Group Pre-Rating Post-Rating

PAGE 58

58 Table 4 4. Study one guava group mean solution ratings and p values for normal weight group Strawberry Peach Raspberry Lychee Guava Pre rating 19.92 13.11 3.65 0.84 4.92 Post rating 14.54 18.57 4.95 1.43 2.92 P value 0.3498 0.1112 0.8126 0.9061 0.1408 Table 4 5. Study one guava group mean solution ratings and p values for overweight group Table 4 6. Study one lychee group mean solution ratings and p values for normal weight group Strawberry Peach Raspberry Lychee Guava Pre rating 17.89 9.33 10.04 4.22 1.04 Post rating 13.22 13.37 17.74 12.11 10.52 P value 0.4045 0.4687 0.1853 0.2714 0.1498 Table 4 7. Study one lychee group mean solution ratings and p values for overweight group Strawberry Peach Raspberry Lychee Guava Pre rating 33.70 17.15 0.05 15.10 1.65 Post rating 26.55 10.75 1.90 10.05 3.05 P value 0.1579 0.4376 0.7672 0.5997 0.8542 Strawberry Peach Raspberry Lychee Guava Pre rating 10.54 6.62 0.08 24.62 12.15 Post rating 12.62 7.62 5.69 1.46 1.69 P value 0.8098 0.9130 0.6774 0.0165 0.2109

PAGE 59

59 Figure 4 5. Study one guava group mean solution ratings for normal weight group with p values indicated above each sample Figure 4 6. Study one guava group mean solution ratings for overweight group with p values indicated above each sample -10.00 -5.00 0.00 5.00 10.00 15.00 20.00 25.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Normal Weight Group Average Solution Ratings: Guava group Pre-Rating Post-Rating -30.00 -25.00 -20.00 -15.00 -10.00 -5.00 0.00 5.00 10.00 15.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Overweight Group Average Solution Ratings: Guava group Pre-Rating Post-Rating 0.3498 0.1112 0.8126 0.1408 0.9061 0.8098 0.9130 0.677 4 0.0165 0.2109

PAGE 60

60 Figure 4 7. Study one lychee group mean solution ratings for normal weight group with p values indicated above each sample Figure 4 8. Study one lychee group mean solution ratings for overweight group with p values indicated above each sample -5.00 0.00 5.00 10.00 15.00 20.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Normal Weight Group Average Solution Ratings: Lychee group Pre-Rating Post-Rating -20.00 -10.00 0.00 10.00 20.00 30.00 40.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Overweight Group Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.4045 0.1853 0.4687 0.2714 0.1498 0.1579 0.4376 0.7672 0.5997 0.8542

PAGE 61

61 Table 4 8. Overall mean ice cream ratings for n ormal weight and overweight BMI groups for s tudy one and study two, including p values for comparison of normal weight to overweight for each exposure Study 1 Study 2 Normal Weight Overweight P Value Normal Weight Overweight P Value N = 64 N = 33 N = 78 N = 26 Exposure 1 42.47 54.15 0.0678 34.81 40.62 0.3613 Exposure 2 N/A N/A N/A 33.33 40.65 0.2164 Exposure 3 N/A N/A N/A 36.15 39.04 0.6601

PAGE 62

62 Table 4 9 Study one guava group mean solution ratings and p values for female group Strawberry Peach Raspberry Lychee Guava Pre rating 23.70 16.18 2.91 8.94 11.18 Post rating 16.03 19.27 5.52 3.36 1.55 P value 0.2299 0.4435 0.2383 0.3954 0.0911 Figure 4 9. Study one guava group mean solution ratings for female group with p values indicated above each sample -15.00 -10.00 -5.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Female Average Solution Ratings: Guava group Pre-Rating Post-Rating 0.2299 0.4435 0.2383 0.3954 0.0911

PAGE 63

63 Table 4 10 Study one guava group mean solution ratings and p values for male group Strawberry Peach Raspberry Lychee Guava Pre rating 5.41 2.18 13.53 3.29 1.71 Post rating 10.18 8.82 4.41 2.29 8.06 P value 0.4728 0.3051 0.2011 0.2250 0.3803 Figure 4 10. Study one guava group mean solution ratings for male group with p values indicated above each sample -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Male Average Solution Ratings: Guava group Pre-Rating Post-Rating 0.2011 0.4728 0.3051 0.3803 0.2250

PAGE 64

64 Table 4 11 Study one lychee group mean solution ratings and p values for female group Strawberry Peach Raspberry Lychee Guava Pre rating 24.80 15.20 14.70 6.57 0.47 Post rating 22.83 16.50 17.00 13.53 8.80 P value 0.7020 0.8262 0.6847 0.3579 0.3028 Figure 4 11. Study one lychee group mean solution ratings for female group with p values indicated above each sample 0.00 5.00 10.00 15.00 20.00 25.00 30.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Female Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.7020 0.8262 0.6847 0.3579 0.3028

PAGE 65

65 Table 4 12 Study one lychee group mean solution ratings and p values for male group Strawberry Peach Raspberry Lychee Guava Pre rating 24.29 8.18 9.94 22.65 0.53 Post rating 11.94 4.77 0.41 16.47 4.77 P value 0.0236 0.6715 0.0948 0.4851 0.4004 Figure 4 12. Study one lychee group mean solution ratings for male group with p values indicated above each sample -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Male Average Solution Ratings: Lychee Group Pre-Rating Post-Rating 0.0236 0.6715 0.0948 0.4851 0.4004

PAGE 66

66 Table 4 13 Study two guava group mean ratings and p values for flavored solutions Strawberry Peach Raspberry Lychee Guava Pre rating 3.17 8.12 6.04 6.87 4.15 Post rating 2.00 0.42 3.21 0.33 0.35 P value 0.1788 0.0107 0.4266 0.0410 0.2389 Table 4 14 Study two lychee group mean ratings and p values for flavored solutions Strawberry Peach Raspberry Lychee Guava Pre rating 8.23 12.33 3.62 14.71 3.79 Post rating 6.15 14.23 10.60 14.94 1.50 P value 0.5706 0.5042 0.0535 0.9402 0.4257 Figure 4 13. Study two guava group mean ratings for flavored solutions with p values indicated above each sample -6 -4 -2 0 2 4 6 8 10 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Average Ratings for Flavored Solutions: Guava Group Pre-Rating Post-Rating 0.0107 0.4266 0.0410 0.1788 0.2389

PAGE 67

67 Figure 4 14. Study two lychee group mean ratings for flavored solutions with p values indicated above each sample Figure 4 15. Study two average ice cream ratings for lychee and guava across three sessions 0 2 4 6 8 10 12 14 16 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Average Ratings for Flavored Solutions: Lychee Group Pre-Rating Post-Rating 0 5 10 15 20 25 30 35 40 45 Lychee Guava Average Rating Ice Cream Sample Study Two Average Ice Cream Ratings Session 1 Session 2 Session 3 0.5042 0.9402 0.5706 0.0535 0.4257

PAGE 68

68 Table 4 15 Study two fruit consumption history Fruit Yes No Strawberry 104 0 Peach 104 0 Papaya 83 21 Raspberry 100 4 Orange 103 1 Blueberry 101 3 Apricot 83 21 Passionfruit 85 19 Lychee 80 24 Mango 101 3 Guava 81 23 Pomegranate 97 7 Blackberry 94 10 Pineapple 102 2 Kiwi 101 3

PAGE 69

69 Figure 4 16. Study two distribution of flavored solution ratings for guava group Figure 4 17 Study two distribution of flavored solution ratings for lychee group 0 2 4 6 8 10 12 14 16 18 -100 -50 0 50 100 Rating Frequency Individual Raw Rating Distribution of Guava Solution Pre and Post Ratings: Guava Group Pre-Rating Post-Rating 0 2 4 6 8 10 12 14 16 -100 -50 0 50 100 Rating Frequency Individual Raw Rating Distribution of Lychee Solution Pre and Post Ratings: Lychee Group Pre-Rating Post-Rating

PAGE 70

70 Table 4 16 Study two guava group mean solution ratings and p values for normal weight group Strawberry Peach Raspberry Lychee Guava Pre rating 4.10 7.91 6.41 8.24 4.33 Post rating 0.43 0.26 4.19 2.26 0.83 P value 0.2698 0.0284 0.5627 0.1161 0.3137 Table 4 17 Study two guava group mean solution ratings and p values for overweight group Strawberry Peach Raspberry Lychee Guava Pre rating 0.70 9.00 4.50 1.10 3.40 Post rating 8.60 3.30 0.90 11.20 1.70 P value 0.4713 0.2210 0.5837 0.1969 0.5712 Table 4 18 Study two lychee group mean solution ratings and p values for normal weight group Table 4 19 Study two lychee group mean solution ratings and p values for overweight group Strawberry Peach Raspberry Lychee Guava Pre rating 1.75 10.50 3.88 8.56 2.00 Post rating 1.94 11.13 9.63 13.50 1.63 P value 0.9703 0.8244 0.1702 0.3844 0.9308 Strawberry Peach Raspberry Lychee Guava Pre rating 11.11 13.14 3.50 17.44 4.58 Post rating 8.03 15.61 11.03 15.58 1.44 P value 0.5254 0.5324 0.1270 0.6160 0.4002

PAGE 71

71 Figure 4 18. Study two guava group mean solution ratings for normal weight group with p values indicated above each sample Figure 4 19. Study two guava group mean solution ratings for overweight group with p values indicated above each sample -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00 8.00 10.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Normal Weight Group Average Solution Ratings: Guava group Pre-Rating Post-Rating -15.00 -10.00 -5.00 0.00 5.00 10.00 15.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Overweight Group Average Solution Ratings: Guava group Pre-Rating Post-Rating 0.0284 0.1161 0.5627 0.2698 0.3137 0.2210 0.5837 0.4713 0.1969 0.5712

PAGE 72

72 Figure 4 20. Study two lychee group mean solution ratings for normal weight group with p values indicated above each sample Figure 4 21. Study two lychee group mean solution ratings for overweight group with p values indicated above each sample 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Normal Weight Group Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Overweight Group Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.6160 0.5324 0.5254 0.1270 0.4002 0.3844 0.8244 0.1702 0.9703 0.9308

PAGE 73

73 Table 4 20 Study two guava group mean solution ratings and p values for female group Strawberry Peach Raspberry Lychee Guava Pre rating 6.45 5.72 3.24 8.52 6.38 Post rating 0.03 1.79 4.62 5.72 2.28 P value 0.3181 0.1361 0.8042 0.5982 0.4150 Table 4 21 Study two guava group mean solution ratings and p values for male group Strawberry Peach Raspberry Lychee Guava Pre rating 0.96 11.13 9.57 4.78 1.35 Post rating 4.57 1.30 1.44 7.96 2.09 P value 0.2961 0.0216 0.042 0 0.0034 0.3689 Table 4 22 Study two lychee group mean solution ratings and p values for female group Strawberry Peach Raspberry Lychee Guava Pre rating 6.49 10.70 1.33 10.79 1.73 Post rating 2.52 12.42 10.24 10.94 3.30 P value 0.4774 0.6859 0.0295 0.9728 0.7096 Table 4 23 Study two lychee group mean solution ratings and p values for male group Strawberry Peach Raspberry Lychee Guava Pre rating 11.26 15.16 12.21 21.53 13.37 Post rating 12.47 17.37 11.21 21.90 9.84 P value 0.6654 0.4081 0.7706 0.9184 0.2429

PAGE 74

74 Figure 4 22. Study two guava group mean solution ratings for female group with p values indicated above each sample Figure 4 23. Study two guava group mean solution ratings for male group with p values indicated above each sample -8.00 -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00 8.00 10.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Female Average Solution Ratings: Guava group Pre-Rating Post-Rating -10.00 -5.00 0.00 5.00 10.00 15.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Male Average Solution Ratings: Guava group Pre-Rating Post-Rating 0.5982 0.3181 0.1361 0.8042 0.4150 0.0216 0.042 0 0.0034 0.2961 0.3689

PAGE 75

75 Figure 4 24. Study two lychee group mean solution ratings for female group with p values indicated above each sample Figure 4 25. Study two lychee group mean solution ratings for male group with p values indicated above each sample -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Female Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.00 5.00 10.00 15.00 20.00 25.00 Strawberry Peach Raspberry Lychee Guava Average Rating Flavored Solution Male Average Solution Ratings: Lychee group Pre-Rating Post-Rating 0.6859 0.9728 0.0295 0.4774 0.7096 0.9184 0.4081 0.6654 0.7706 0.2429

PAGE 76

76 CHAPTER 5 CONCLUSION The main objective of this study was to better understand human food choice in regards to the mechanisms responsible for the development of food liking. This objective was met by investigating if hedonic ratings of two novel flavors would increase after being paired with an innately positive foo d matrix. Results indicate there is the potential for flavors to become liked more after being associated with an innately positive food matrix. Variations in hedonic ratings were present in both Study one and Study two demonstrating increased ratings of experimental flavors. Most promising was the near significant increase for the guava flavored solution of the guava group in Study one. However, the hypothesis was rejected th at the hedonic rating of flavors will increase after being paired with an innately positive food matrix (ice cream) There were no significant increases (p<0.05) in the ratings of the experimental flavors for Study one or Study two. Additionally, no clear differences were observed when data were sorted by gender and BMI. However, the changes in hedonic ratings observed are encouraging for future research in this area. Specifically, the consistency of the observed variations in hedonic ratings between both s tudies is promising, as it indicates both strengths and weakness es of this particular experimental design. Integrating alterations to the experimental design based on the possible sources of error observed in this project would be useful for further resear ch in this area. Though the results show some increases in ratings, the lack of significant increases begs a discussion of future directions. Most notable from the results of these studies are the considerably low ratings for the flavored sugar solutions. As mentioned previously, these samples were formulated to create samples which are simple and isolate a single

PAGE 77

77 flavor. This goal was achieved, but the samples were disliked strongly from a subset of panelists. Although the solutions were sweetened to 5% su crose, it is apparent the concentrated flavors added to the solution were not pleasant singularly. A sugar solution was therefore not the most natural food matrix for isolating singular flavors. Perhaps a whipped dairy product, unsweetened yogurt, or anoth er similar palatable matrix would have served the purpose better. Overall, the food matrix for addition of the isolated flavors would need to be a relatively neutral or positively received sample. Beyond the personal likeability of the samples, hunger may play a vital role in reinforcing a flavor. Future studies in this area would benefit by incorporating this as a variable to discover its importance for driving flavor liking and how our body responds to food when hungry in regards to flavor learning. Addit ionally, the novelty of flavors to panelists is a variable which must be considered when attempting to drive the liking of a flavor. Although the clearly indicate t he importance of this variable. However, subjects most likely should be unfamiliar with a flavor, as changing the opinion of a previously familiar flavor would be a formidable task. Finally, the variability among sample flavors is a variable which must be considered. The studies in this project focused on fruit flavors, which are fairly similar. Future research in this area may need to consider utilizing samples which are non overlapping. In other words, samples which differ greatly in flavor should be expl oited (eg. peanut butter, fruit, vegetable, meat, bread, etc.). The above mentioned variables should be considered for future research in this area, as these studies are promising for developing a better understanding of flavor learning. Above all investi gating the driving factors of flavor liking would provide further

PAGE 78

78 information about the olfactory system and how our brain processes flavors to create memory templates. Furthermore, this research and future rese arch in this area can lead to an overarching understanding of acceptability, preference, and affective responses to foods in humans.

PAGE 79

79 APPENDIX A QUESTIONNAIRE Please complete the form below. If you pass the screening criteria, you will receive an email b y the evening of Friday, September 14 Required First name* ____________________ Last name* ____________________ E mail* ____________________ Question 1* Which of the following fruits have you ever consum ed in any form (including raw fruit, juice, juice blend)? Yes No Strawberry Peach Papaya Raspberry Orange Blueberry Apricot Passion fruit Lychee Mango Guava Pomegranate Blackberry Pineapple Kiwi Question 2* Is there anything else we should know (food allergies, food restrictions, etc.)? ________________________________________________________________ ________________________________________________________________ ________________________________________________________________

PAGE 80

80 APPENDIX B HGLMS

PAGE 81

81 APPENDIX C IRB FORM Fresh & Processed Foods Taste Panel Consent Form I, Charles Sims (Professor, Food Science and Human Nutrit ion Department), am requesting your consent to participate in the sensory testing of fresh and processed foods. You will be given between 1 and 10 samples and will be asked to evaluate the different attributes of the foods. We simply want your opinion of these products. This should take about 5 10 minutes of your time. Your name will not appear on the ballot that you fill out to keep your responses confidential to the extent provided by the law. These food products are produced by commercial companies as part of their research and development activities and are very similar to commercial foods available in retail markets. You will be verbally informed of the exact type of food product and if the product contains aspartame (Nutrasweet). All ingredient s in these products are approved for use in foods. Products will be served at temperatures generally associated with the food products, including cool temperatures for some and heated and/or room temperature for others. We will handle and present the pro ducts to you in a sanitary manner. You will be asked to evaluate the products in a private booth, and water to rinse your mouth and crackers will be provided. To the best of our knowledge, there are no known risks associated with tasting these products, nor are there any immediate benefits. Your participation is voluntary, and you are free to withdraw at any time. You will be given a reward for your participation, up to $10.00 per session attended. ts can be directed to the UFIRB Office, PO Box 112250, University of Flori da, Gainesville, FL 32611 2250. If you need further information about this project, please contact me at 392 1991 or come by my office at Room 130 Aquatic Foods Pilot Plant (Building 120) ____________ Principa Date I have read the procedure described above and have been offered a copy of this description, and I agree to participate in the procedure. Date:

PAGE 82

82 APPENDIX D COMPUSENSE TEST BALLOT Today's Sample: Flavored Sugar Solutions To start the test, click the Continue button below: CONTINUE Panelist Registration Number: ________

PAGE 83

83 Please indicate your gender. o Male o Female Please enter your age. Age ______________ Pleas e enter your height (For example: If you are 5 feet and 3 inches in height, enter 503). Height ______________ Please enter your weight in pounds. Wei ght ______________ What is your ethnic background? o Hispanic o Non Hispanic Which of the following best describes you? o Asian/Pacific Islander o Black or African American o White or Caucasian o Native American, Alaska Native, Aleutian o Other Have you ever suffered from middle ear infections? o No o Yes, but not serious o Yes, required antibiotics more than once o Yes, required tubes in ears

PAGE 84

84 S C A L E 1 1. Now, please take a few minutes to identify the strongest LIKING (i.e., pleasure) of any kind that you have ever experienced. 2. Once you have identified your strongest LIKING experienced, please write it down on the paper provided and type it in on the next screen. 3. Please remember to use the strongest liking that you've identified, and written down as the top of your scale (100). Please click on the 'Continue' button below. Question # 1. Please type the strongest LIKING OF ANY KIND YOU'VE EXPERIENCED in the space below and remember that this sensatio n will be 100 on your scale ( S C A L E 1 ). ____________________________________________________________________________________ ____________________________________________________________________________________ ___________________________________ _________________________________________________

PAGE 85

85 S C A L E 1 1. Now, please take a few minutes to identify the strongest DISLIKING (i.e., displeasure) of any kind that you have ever experienced. 2. Once you have identified your strongest DISLIKING experienced, please write it down on the paper provided and type it in on the next screen. 3. Please remember to strongest disliking that you've identified, and written down as the bottom of your scale ( 100). Please click on the 'Conti nue' button below. Question # 2. Please type the strongest DISLIKING OF ANY KIND YOU'VE EXPERIENCED in the space below and remember that this sensation will be 100 on your scale ( S C A L E 1 ). ____________________________________________________________________________________ ____________________________________________________________________________________ ____________________________________________________________________________________

PAGE 86

86 S C A L E 1 On the line scale, 100 indicates the most intense liking (i.e., pleasure) you have ever experienced (no matter what the source). Similarly, 100 indicates the opposite: the most intense disliking you have ever experienced. Neutral is indicated by 0. Please use your 100 and 100 (written on the paper provided) to answer the following questions. Please click the 'Continue' button below.

PAGE 87

87 Question # 3 Please rate the experiences listed below from memory using your S C A L E 1. Listening to your Favorite music Strongest Neutral Strongest Dislike Like 100 0 100 Spending time with your loved ones Strongest Neutral Strongest Dislike Like 100 0 100 The death of a loved one Strongest Neutral Strongest Dislike Like 100 0 100 The most intense anger you've experienced Strongest Neutral Strongest Dislike Like 100 0 100 The most nervous you have ever been Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 88

88 The most intense annoyance you've experienced Strongest Neutral Strongest Dislike Like 100 0 100 Eating your favorite food Strongest Neutral Strongest Dislike Like 100 0 100 Eating your least favorite food Strongest Neutral Strongest Dislike Like 100 0 100 Proudest you've ever been of accomplishing a specific goal Strongest Neutral Strongest Dislike Like 100 0 100 The most Enthusiastic you've ever been about a hobby Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 89

89 The Shyest you've ever been Strongest Neutral Strongest Dislike Like 100 0 100 The most Amused you have ever been by an anecdote Strongest Neutral Strongest Dislike Like 100 0 100 The most Inspired you have ever been by a lecture. Strongest Neutral Strongest Dislike Like 100 0 100 The most Disgusted you have ever been by a specific food Strongest Neutral Strongest Dislike Like 100 0 100 The Best tomato you've ever tasted Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 90

90 The Worst tomato you've ever tasted Strongest Neutral Strongest Dislike Like 100 0 100 PLEASE LIFT THE WINDOW TO RECEIVE YOUR SAMPLES.

PAGE 91

91 You are now ready to taste sample <> Take a bite of cracker and a sip of water to rinse your mouth. WHEN ANSWERING ANY QUESTION, MAKE SURE THE NUMBER ON THE CUP MATCHES THE NUMBER ON THE MONITOR. Please click on the 'Continue' button below. Question # 1 Sample <> Please use your S C A L E 1 to rate your overall liking of Sample <>. OVERALL LIKING Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 92

92 You are now ready to taste sample <> Take a bite of cracker and a sip of water to rinse your mouth. WHEN ANSWERING ANY QUESTION, MAKE SURE THE NUMBER ON THE CUP MATCHES THE NUMBER ON THE MONITOR. Please click on the 'Continue' button below. Question # 1 Sample <> Please use your S C A L E 1 to rate your overall liking of Sample <>. OVERALL LIKING Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 93

93 You are now ready to taste sample <> Take a bite of cracker and a sip of water to rinse your mouth. WHEN ANSWERING ANY QUESTION, MAKE SURE THE NUMBER ON THE CUP MATCHES THE NUMBER ON THE MONITOR. Please click on the 'Continue' button below. Question # 1 Sample <> Please us e your S C A L E 1 to rate your overall liking of Sample <>. OVERALL LIKING Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 94

94 You are now ready to taste sample <> Take a bite of cracker and a sip of water to rinse your mouth. WHEN ANSWERING ANY QUESTION, MAKE SURE THE NUMBER ON THE CUP MATCHES THE NUMBER ON THE MONITOR. Please click on the 'Continue' button below. Question # 1 Sample <> Please us e your S C A L E 1 to rate your overall liking of Sample <>. OVERALL LIKING Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 95

95 You are now ready to taste sample <> Take a bite of cracker and a sip of water to rinse your mouth. WHEN ANSWERING ANY QUESTION, MAKE SURE THE NUMBER ON THE CUP MATCHES THE NUMBER ON THE MONITOR. Please click on the 'Continue' button below. Question # 1 Sample <> Please use your S C A L E 1 to rate your overall liking of Sample <>. OVERALL LIKING Strongest Neutral Strongest Dislike Like 100 0 100

PAGE 96

96 Please Lift your window! Today's test is complete. You must return for the next session! Check your email for your scheduled time.

PAGE 97

97 LIST OF REFERENCES Ackroff K, Sclafani A. 1991. Flavor preferences conditioned by sugars: Rats learn to prefer glucose over fructose. Physiology & Behavior 50(4):815 24. Adler E, Hoon MA, Mueller KL, Chandrashekar J, Ryba NJP, Zuker CS. 2000. A novel family of mammalian taste receptors. Cell 100(6):693 702. Auvray M, Spence C. 2008. The multisensory perception of flavor. Consciousness and Cognition 17(3):1016 31. Bartoshuk L. 2010. Hedonic gLMS: a new scale that permits valid hedonic comparisons. International Citrus & Beverage Conference. Bartoshuk LM. 1975. Taste mixtures: Is mixture suppression related to compression? Physiology & Behavior 14(5):643 9. Bartoshuk LM, Duffy VB, Fast K, Green BG Prutkin J, Snyder DJ. 2003. Labeled scales (e.g., category, Likert, VAS) and invalid across group comparisons: What we have learned from genetic variation in taste. Food Quality and Preference 14(2):125 38. Bartoshuk LM, Duffy VB, Green BG, Hoffman H J, Ko CW, Lucchina LA, Marks LE, Snyder DJ, Weiffenbach JM. 2004. Valid across group comparisons with labeled scales: the gLMS versus magnitude matching. Physiology & Behavior 82(1):109 14. Bartoshuk LM, Fast K, Snyder DJ. 2005. Differenc es in our sens ory w orlds: Invalid comparisons with labeled s cales. Current Directions in Psychological Science 14(3):122 5. Bernstein I. 1978. Learned taste aversions in children receiving chemotherapy. Science 200(4347):1302 3. Birch LL. 1999. Development of food preferences. Annual Review of Nutrition 19(1):41 62. Cornier M A, Salzberg AK, Endly DC, Bessesen DH, Tregellas JR. 2010. Sex based differences in the behavioral and neuronal responses to food. Physiology & Behavior 99(4):538 43. Dairy Mix, Inc.: Prod ucts. St. Petersburg, FL: Dairy Mix, Inc.; 2006 [Accessed 2013 May 23] Available from http://www.dairymix.com Drewnowski A. 1997. Taste Preferences and Food Intake. Annual Review of Nutrition 17:237 53.

PAGE 98

98 Gibson EL, Wainwright CJ, Booth DA. 1995. Disgu ised protein in lunch after low protein breakfast conditions food flavor preferences dependent on recent lack of protein intake. Physiology & Behavior 58(2):363 71. Green BG, Dalton P, Cowart B, Shaffer G, Rankin K, Higgins J. 1996. Evaluating the Chemical Senses 21:323 34. Green BG, Shaffer GS, Gilmore MM. 1993. Derivation and evaluation of a semantic scale of oral sensation magnitude with apparent ratio properties. Chemical Senses 18(6):683 702. Jinks A, Laing D. 1999a. A limit in the processing of components in odour mixtures. Perception 28:395 404. Jinks A, Laing DG. 1999b. Temporal processing reveals a mechanism for limiting the capacity of humans to analyze odor mixtures. Cognitive Brain Research 8(3):311 25. Johnson SL, McPhee L, Birch LL. 1991. Conditioned preferences: Young children prefer flavors associated with high dietary fat. Physiology & Behavior 50(6):1245 51. Jones LV, Peryam DR, Thurstone LL. 1955. Development of a scale for measuring soldiers' food preferences. Journal of Food Science 20(5):512 20. Laing DG, Francis GW. 1989. The capacity of humans to identify odors in mixtures. Physiology & Behavior 46(5):809 14. Laing DG, Link C, Jinks AL, Hutchinson I. 2002. The limited capacity of humans to identify the components of taste mixtures and taste odour mixtures. Perception (London) 31(5):617. Lancet D. 1986. Vertebrate olfactory reception. Annual Review of Ne uroscience 9(1):329 55. Lawless HT, Heymann H. 1999. Sensory Evaluation of Food: Principles and Practices. New York, NY: Springer Science + Buisness Media, LLC. Livermore A, Laing DG. 1996. Influence of training and experience on the perception of mul ticomponent odor mixtures. Journal of Experimental Psychology: Human Perception and Performance 22(2):267 77. Marshall K, Laing DG, Jinks AL, Effendy J, Hutchinson I. 2005. Perception of temporal order and the identification of components in taste mixtu res. Physiology & Behavior 83(5):673 81.

PAGE 99

99 Meilgaard MC, Civille GV, Carr BT. 2007. Sensory Evaluation Techniques, 4 ed. Boca Raton, FL: CRC Press. Meiselman HL. 1993. Critical evaluation of sensory techniques. Food Quality and Preference 4(1 2):33 40. Methven L, Langreney E, Prescott J. 2012. Changes in liking for a no added salt soup as a function of exposure. Food Quality and Preference 26(2):135 40. Mori K, Nagao H, Yoshihara Y. 1999. The olfactory bulb: Coding and processing of odor molecule inf ormation. Science 286(5440):711 5. Myers KP, Sclafani A. 2006. Development of learned flavor preferences. Developmental Psychobiology 48(5):380 8. National Institute of Diabetes and Digestive and Kidney Diseases. National Institutes of Health. Bethesda, MD: Weight and Waist Measurement: Tools for Adults. Weight control Information Network; 2012 [Accessed 2013 March 22] Available from: http://win.niddk.nih.gov Nelson GM. 1998. Biology of taste buds and the clinical problem of taste loss. The Anatomical Record (New Anat.) 253:70 8. Rissanen A, Hakala P, Lissner L, Mattlar CE, Koskenvuo M, Ronnemaa T. 2002. Acquired preference especially for dietary fat and obesity: A study of weight discordant monozygotic twin pairs. International Journal of Obesity and Related Metabolic Disorders 26(7):973 7. Attention, Perception, & Psychophysics 31(4):397 401. Schutz HG, Cardello AV. 2001 A Labeled Affective Magnitude (LAM) scale for assessing food liking/disliking. Journal of Sensory Studies 16(2):117 59. Sclafani A. 1997. Learned controls of ingestive behaviour. Appetite 29(2):153 8. Shepherd G. 1994. Discrimination of molecular sign als by the olfactory receptor neuron. Neuron 13(4):771 90. Shepherd GM. 2004. The human sense of smell: Are we better than we think? PLoS Biol 2(5):e146. Shepherd GM. 2005. Outline of a theory of olfactory processing and its relevance to humans. Chemi cal Senses 30:I3 I5.

PAGE 100

100 Shepherd GM. 2006. Smell images and the flavour system in the human brain. Nature 444(7117):316 21. Snyder D, Puentes L, Sims C, Bartoshuk L. 2008. Building a better intensity scale: Which labels are essential? Chemical Senses 33(8 ):S142. Stevens SS. 1957. On the psychophysical law. Psychological Review 64(3):153 81. Stevens SS. 1960. The psychophysics of sensory function. American Scientist 48(2):226 53. Vassar R, Chao SK, Sitcheran R, Nuez JM, Vosshall LB, Axel R. 1994. Topog raphic organization of sensory projections to the olfactory bulb. Cell 79(6):981 91. Williams EJ. 1949. Experimental designs balanced for the estimation of residual effects of treatments. Australian Journal of Scientific Research 2:149 68. Wilson DA, Stevenson RJ. 2003. The fundamental role of memory in olfactory perception. Trends in Neurosciences 26(5):243 7. Yeomans MR 2008. Learning and hedonic contributions to human o besity. In: Blass EM, editor. Obesity: Causes, Mechanisms, Prevention, and Tr eatment. Sunderland, MA: Sinauer Associates, Inc. p. 211 42. Zellner DA, Rozin P, Aron M, Kulish C. 1983. Conditioned enhancement of human's liking for flavor by pairing with sweetness. Learning and Motivation 14(3):338 50. Zhao H, Ivic L, Otaki JM, Ha shimoto M, Mikoshiba K, Firestein S. 1998. Functional expression of a mammalian odorant receptor. Science 279(5348):237 42.

PAGE 101

101 BIOGRAPHICAL SKETCH Elizabeth Gardner was born in Rockledge, Florida to Michael and Suzanne Gardner. She grew up in Cocoa, Florida and attended Cocoa High School. Following graduation from high school, Elizabeth attended the University of Florida from 2 007 to 2011, graduating with a bachelor of s cience in f ood s cience and h uman n utrition, with a specialization in f ood s cience. She extended her stay at the University of Florida for two years to receive her m aster of s cience in f ood s cience degree in August 2013. Throughout her academic career, Elizabeth was involved in the University of Florida Gator Marching Band and University of Florida Concert Bands. As an undergraduate student she was an active member of Sigma Kappa sorority and participated in