SLNOM: Exploring the sound of mastication as a behavioral change strategy for rapid eating regulation

Abstract

Rapid eating is linked to numerous health problems, such as obesity and gastritis. In this study, we explore the possibility of using mastication sound as a novel behavior change strategy to subtly regulate rapid eating behavior. In particular, we present SLNOM, a system that can automatically detect chewing behavior using a convolutional neural network (CNN) model, and slow down the playback speed of real-time mastication sounds to implicitly modify eating behavior. Two empirical studies have been conducted to determine: 1) the threshold of sound volume and speed without user perception; and 2) the feasibility and effectiveness of SLNOM in changing eating behavior using a Wizard of Oz study. The result indicated that manipulation of chewing sound could modulate eating rate, bite size without cognitive and behavioral effort. We discussed how cognitive science could explain these findings and suggested how future eating interventions can be designed to take advantage of current exploration.