The development and testing of the TTU food cue reactivity image bank.

Background/objectives: Measuring food cue reactivity (FCR) is essential to understanding human ingestion in behavioral and neuroimaging studies. An image bank that can delineate the truly food-specific FCR response from spurious noise driven primarily by incidental characteristics (e.g., visual properties) of images is a useful addition to the literature. This study sought to develop and test the performance of the TTU Food Cue Reactivity Image Bank that matched 252 image pairs on visual characteristics (i.e., shape, color, visual complexity, and size), and to establish the appeal ratings of the images for use in future applications.

Subjects/methods: The TTU Food Cue Reactivity Image Bank was initially created, and subsequently evaluated by independent raters. Then, 151 individuals participated in a Qualtrics survey to determine the similarity of image pairs on relevant dimensions (i.e., shape, color, visual complexity, and size) and establish appeal ratings for food and non-food images.

Results: Inter-rater agreement was tested using intraclass correlation coefficients (ICC), which revealed very high agreement among raters for all similarity measures (shape ICC = 0.98; color ICC = 0.97; visual complexity ICC = 0.96; size ICC = 0.96; appeal ICC = 0.93).

Conclusions: The high agreement among raters on the visual characteristics and appeal ratings of the images increases confidence that food-cue-reactivity observed is based on intended "image type" distinctions (i.e., food vs. object) and not incidental visual features.