The perception of sweet, bitter and fatty acid taste and sensitivity to fat by mouthfeel and olfaction : associations with dietary intake, eating behaviour and obesity in premenopausal women : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Nutritional Science at Massey University, Albany, New Zealand

Figures are re-used with permission. The PROMISE paper (pp. 325-340) published in JMIR Research Protocols (http://www.researchprotocols.org/2019/8/e14529) is an open access article distributed under a Creative Commons Attribution License 4.0.Background: Individual variability in taste perception may influence diet, possibly modifying eating behaviour and long-term food choice. Research into taste perception and weight status, dietary intake, eating behaviour and endocrine regulators of metabolic health could provide new important insights. Taste perception may be modifiable, and as such may be a target for future intervention strategies which may have the potential to prevent or treat obesity. Objectives: The aims of this study were to determine associations between (1) fatty acid taste, olfaction, mouthfeel of fat, dietary intake, eating behaviour and body mass index (BMI), and (2) associations between taste perception of glucose (sweet taste), quinine (bitter taste) and milk (fat perception) with body composition and hormonal adiposity signals (fasting insulin and leptin), dietary intake and eating behaviour. Design: For the first cross-sectional study, 50 premenopausal women assessed oleic acid taste and olfaction thresholds. BMI was calculated from weight (kg) and height (m). Dietary intake and eating behaviour were evaluated using a food frequency and three-factor eating questionnaire (TFEQ), respectively. Binomial regression analysis was used to model fatty acid taste and olfaction data and fatty acid taster status was determined (hypersensitive, n= 22; hyposensitive, n= 28). For the second cross-sectional study (the PROMISE study), 351 premenopausal Pacific and New Zealand (NZ) European women were recruited and stratified by non-obese and obese groups. Suprathreshold intensity, hedonic liking, and discrimination of taste by a ranking task were measured using a range of concentrations of sweet, bitter, and fat solutions. Participants were classified as likers or dislikers for each tastant using a hierarchical cluster analysis. Body fat (BF) was quantified by dual x-ray absorptiometry. Total energy and macronutrient intake were assessed using a 5-day estimated food record and eating behaviour was assessed by TFEQ. Socioeconomic status was measured by deprivation index (NZDep2013). Logistic and linear regression analyses were used to analyse study outcomes and to adjust for potential confounders (socioeconomic position, age, etc). Both studies were undertaken in Auckland, NZ. Outcomes: The results of the first study showed taste and olfactory detection for oleic acid were positively correlated (r= 0.325; P35 BF% (adjusted, OR 2.9, P35 BF% compared to sweet dislikers (adjusted, OR 2.1, P35 BF% in comparison to bitter dislikers (adjusted, OR 0.4, P< 0.01). Having higher fasting plasma leptin concentration significantly increased the likelihood of being a sweet liker in Pacific and in NZ European combined (adjusted; OR 1.7, P< 0.05), but in NZ European women, the likelihood of this was further increased (adjusted; OR 3.6, P< 0.001). Higher fasting plasma insulin concentration also increased the likelihood of being a sweet liker (adjusted, OR 1.7, P< 0.05). New Zealand European sweet likers had a significantly higher intake of carbohydrates, all sugars and starch (199.4 ± 51.1, 87.9 ± 27.4 and 111.1 ± 34.6 g/day) when compared to sweet taste dislikers (165.9 ± 48.7, 71.4 ± 25.2 and 94.1 ± 34.7 g/day, P< 0.001, P< 0.01, and P< 0.01, respectively), however, this comparison was not significant in Pacific women. NZ European women in the sweet likers group had an increased disinhibited eating behaviour score (P< 0.01). Conclusions: Fatty acid taste perception was found to be associated with olfaction, eating behaviour and body composition. The findings from the PROMISE study have shown differences in sweet taste perception in relation to adiposity which is further associated with circulating plasma leptin and insulin concentrations. Sweet taste perception was associated with dietary intake and eating behaviour in NZ European women. The taste-diet associations observed in NZ European women were not observed in Pacific women. Therefore, population groups with lower metabolic disease risk may have dissimilar taste-diet associations compared to those with a higher metabolic disease risk. Taste perception is a promising target for future weight-loss and intervention strategies due to demonstrating links with dietary intake, eating behaviour and body composition