Using participant ratings to construct food image paradigms for use in the Australian population - A pilot study

In human research, images of food are often used as cues in place of real foods. To elicit anticipatory responses in targeted populations (e.g. prompting changes in metabolic hormones, invoking food cravings), cultural differences and population norms with regard to food preferences need to be considered. This pilot study aimed to construct two image paradigms (healthy vs. hyperpalatable foods) for experimental use within the Australian population. A dataset of 200 images (from the licenced database Food-pies and Internet sources), representative of healthy and hyperpalatable foods commonly consumed in Australia, was compiled by research dietitians. Ten male and female adults volunteered to view the images. Participants categorised each image as either healthy food or 'junk food' (Le. hyperpalatable food), and rated each image according to three criteria: 1) familiarity of the food displayed; 2) recognisability of the food; and 3) appetisingness of the food. Overall, agreement with a priori categories was high for both healthy and hyperpalatable food images, 87.3% and 87.7% respectively. The food images with the lowest overall ratings (score < 7 out of possible 9) were removed from the dataset and the final paradigms each contain 75 images. The healthy food paradigm contains foods from the five core food groups (fruit, vegetables, grains and cereals, meat and meat alternatives, dairy foods), and the hyperpalatable food paradigm contains non-core foods (sweet and savoury discretionary choice foods). The paradigms represent a broad range of commonly consumed foods that will be relevant for prospective projects utilising food cues in Australian adults.