Dynamic Context Sensory Testing-A Proof of Concept Study Bringing Virtual Reality to the Sensory Booth

Eating is a multimodal experience. When we eat, we perceive not just the taste and aroma of foods, but also their visual, auditory, and tactile properties, as well as sensory input from our surroundings. Foods are commonly tested within a sensory booth, designed specifically to limit such input. Foods are not commonly experienced in such isolation, but alongside this context, which can alter how a food is perceived. In this study, we show that the sensory properties of food can be altered by changing the environment it is consumed in, using virtual reality (VR) to provide an immersive, dynamic context to the eating experience. The purpose of this project was to develop an affordable and easy-to-implement methodology for adapting VR technology to sensory evaluation, without prohibitive amounts of expensive equipment or specialized programming knowledge. Virtual environments were formed by processing custom-recorded 360 degree videos and overlaying audio, text, sensory scales, and images to simulate a typical sensory evaluation ballot within the VR headset. In a pilot test, participants were asked to taste 3 identical blue cheese samples in 3 virtual contexts-a sensory booth, a park bench, and a cow barn. Respondents rated their liking of the sample, as well as its saltiness, and pungency, attributes either reflective of one context (pungency in the barn), or presumably unrelated (saltiness). Panelists duly rated the sample's flavor as being more pungent when consumed in the barn context. These results provide proof of concept for VR in applied sensory studies, providing an immersive context to a sensory test while remaining in place.

We consume foods in environments that can "spill over" into our perceptions of the food. Thus, we consider some foods "unsuitable" for certain settings, with others deemed more suitable for this locale. This has been studied for many years as sensory "context," with written descriptions, pictures, or videos of such contexts. We present a method generating virtual reality contexts without any specist programming knowledge, for a few hundred dollars. In an accompanying pilot test, perception of a sample was significantly influenced by the VR context in which it was delivered.