Related Experiment Video
Updated: Sep 23, 2026
Control of Eating Behavior Using a Novel Feedback System
Published on: May 8, 2018
Context-aware digital nutrition and abdominal acoustic feedback in Taiwanese university students: a 12-week
Chuanchia Wang1, Chen Han2, Jwoshiun Sun1
1Department of Electronic Engineering, National Taipei University of Technology, Taipei, Taiwan.
Background:
Most digital nutrition tools rely on retrospective meal logging and generic reminders. AbSound adds abdominal acoustic cues intended to support meal-timing reflection and dietary self-regulation. These cues were treated as meal-relevant behavioral proxies, not physiological markers of hunger, satiety, or gastrointestinal motility. We evaluated the platform among Taiwanese university students.
Materials And Methods:
Phase 1 was a 12-week single-arm field intervention with 200 participants completing follow-up. Outcomes included engagement, meal-logging completeness, fruit and vegetable intake, dietary self-regulation, daily steps, WHOQOL-BREF, and BMI. Phase 2 was a separate 4-week open-label randomized feasibility pilot (N = 32; 16 per group) comparing AbSound with standard tracking. Analyses emphasized within-participant trajectories, feasibility estimates, exploratory between-group changes, and agreement with composite meal-state/acoustic-event labels.
Results:
In Phase 1, fruit and vegetable intake, dietary self-regulation, daily steps, and quality of life improved over time, whereas BMI remained stable. Because Phase 1 had no control group, these findings represent temporal associations. Active-day engagement was 78%, meal-logging completeness was 48.5%, and both declined across the semester. In Phase 2, participants used AbSound for 9.35 hours/day across 25.81 of 28 days and acknowledged 78.54% of alerts. Exploratory between-group change estimates directionally favored AbSound for late-night snacking, eating speed, diet quality, and eating self-regulation, but the pilot was not powered for efficacy. The acoustic model achieved 87.2% accuracy and a macro-F1 of 82.8% across 1,207 labeled events; these metrics reflect agreement with composite event labels rather than construct validation of hunger, satiety, or gut motility.
Conclusions:
The findings support the short-term feasibility of context-aware digital nutrition and abdominal acoustic feedback as behavioral self-regulation supports. Larger randomized trials should evaluate efficacy, time-lock acoustic events to validated appetite ratings, use participant-disjoint acoustic validation, and assess longer-term behavioral and clinical outcomes.
