Neural responses to oral administration of erythritol vs. sucrose and sucralose explain differences in subjective liking ratings

Aleksandra Budzinska,Fabienne Teysseire,Emilie Flad,Patrick Dupont,Bettina Wölnerhanssen,Anne Christin Meyer-Gerspach,Lukas Van Oudenhove,Nathalie Weltens

DOI: https://doi.org/10.1016/j.appet.2024.107422

IF: 5.016

2024-09-01

Appetite

Abstract:Introduction: High sugar intake is associated with many chronic diseases. However, non-caloric sweeteners (NCSs) might fail to successfully replace sucrose due to the mismatch between their rewarding sweet taste and lack of caloric content. The natural NCS erythritol has been proposed as a sugar substitute due to its satiating properties despite being non-caloric. We aimed to compare brain responses to erythritol vs. sucrose and the artificial NCS sucralose in a priori taste, homeostatic, and reward brain regions of interest (ROIs). Methods: We performed a within-subject, single-blind, counterbalanced fMRI study in 30 healthy men (mean ± SEM age:24.3 ± 0.8 years, BMI:22.3 ± 0.3 kg/m2). Before scanning, we individually matched the concentrations of both NCSs to the perceived sweetness intensity of a 10% sucrose solution. During scanning, participants received 1 mL sips of the individually titrated equisweet solutions of sucrose, erythritol, and sucralose, as well as water. After each sip, they rated subjective sweetness liking. Results: Liking ratings were significantly higher for sucrose and sucralose vs. erythritol (both pHolm = 0.0037); water ratings were neutral. General Linear Model (GLM) analyses of brain blood oxygen level-depended (BOLD) responses at qFDR<0.05 showed no differences between any of the sweeteners in a priori ROIs, but distinct differences were found between the individual sweeteners and water. These results were confirmed by Bayesian GLM and machine learning-based models. However, several brain response patterns mediating the differences in liking ratings between the sweeteners were found in whole-brain multivariate mediation analyses. Both subjective and neural responses showed large inter-subject variability. Conclusion: We found lower liking ratings in response to oral administration of erythritol vs. sucrose and sucralose, but no differences in neural responses between any of the sweeteners in a priori ROIs. However, differences in liking ratings between erythritol vs. sucrose or sucralose are mediated by multiple whole-brain response patterns.